EURL-GMFF-CT-0117final CTR

EURL GMFF Proficiency Test

Report CT 0217

Determination of GM

Soybean 40-3-2 in Chicken

Feed and Soybean Flour

Broothaerts W Beaz Hidalgo R Corbisier

P Cordeiro F Dimitrievska B Emteborg

H Maretti M Robouch P Emons H

EUR 28873 EN

This publication is a Technical report by the Joint Research Centre (JRC) the European Commissionrsquos science

and knowledge service It aims to provide evidence-based scientific support to the European policymaking

process The scientific output expressed does not imply a policy position of the European Commission Neither

the European Commission nor any person acting on behalf of the Commission is responsible for the use that

might be made of this publication

Contact information European Union Reference Laboratory for GM Food and Feed European Commission Joint Research Centre Directorate F ndash Health Consumers amp Reference Materials Via E Fermi 2749 I-21027 Ispra (VA) Italy Email JRC-EURL-GMFF-CTeceuropaeu Tel +39 0332 78 9952 JRC Science Hub httpseceuropaeujrc JRC109028 EUR 28873 EN

PDF ISBN 978-92-79-76303-8 ISSN 1831-9424 doi 102760461562

Date 20 November 2017

Luxembourg Publications Office of the European Union 2017 copy European Union 2017 Reuse is authorised provided the source is acknowledged The reuse policy of European Commission documents is regulated by Decision 2011833EU (OJ L 330 14122011 p 39) For any use or reproduction of photos or other material that is not under the EU copyright permission must be sought directly from the copyright holders

How to cite this report Broothaerts W Beaz Hidalgo R Corbisier P Cordeiro F Dimitrievska B Emteborg H Maretti M Robouch P Emons H EURL GMFF Proficiency Test Report CT 0217 Determination of GM Soybean 40-3-2 in Chicken Feed and Soybean Flour EUR 28873 EN Publications Office of the European Union Luxembourg 2017 ISBN 978-92-79-76303-8 doi 10-2760461562

All images copy European Union 2017

This publication is a Technical report by the Joint Research Centre (JRC) the European Commissionrsquos science

and knowledge service It aims to provide evidence-based scientific support to the European policymaking

process The scientific output expressed does not imply a policy position of the European Commission Neither

the European Commission nor any person acting on behalf of the Commission is responsible for the use that

might be made of this publication

Contact information European Union Reference Laboratory for GM Food and Feed European Commission Joint Research Centre Directorate F ndash Health Consumers amp Reference Materials Via E Fermi 2749 I-21027 Ispra (VA) Italy Email JRC-EURL-GMFF-CTeceuropaeu Tel +39 0332 78 9952 JRC Science Hub httpseceuropaeujrc JRC109028 EUR 28873 EN

PDF ISBN 978-92-79-76303-8 ISSN 1831-9424 doi 102760461562

Date 20 November 2017

Luxembourg Publications Office of the European Union 2017 copy European Union 2017 Reuse is authorised provided the source is acknowledged The reuse policy of European Commission documents is regulated by Decision 2011833EU (OJ L 330 14122011 p 39) For any use or reproduction of photos or other material that is not under the EU copyright permission must be sought directly from the copyright holders

How to cite this report Broothaerts W Beaz Hidalgo R Corbisier P Cordeiro F Dimitrievska B Emteborg H Maretti M Robouch P Emons H EURL GMFF Proficiency Test Report CT 0217 Determination of GM Soybean 40-3-2 in Chicken Feed and Soybean Flour EUR 28873 EN Publications Office of the European Union Luxembourg 2017 ISBN 978-92-79-76303-8 doi 10-2760461562

All images copy European Union 2017

EURL GMFF Proficiency test report CT

0217

Determination of GM Soybean 40-3-2

in Chicken Feed and Soybean Flour

Broothaerts W Beaz Hidalgo R Corbisier P Cordeiro F Dimitrievska B Emteborg H Maretti M Robouch P Emons H

268-PT Accredited by the

Belgian Accreditation Body (BELAC)

Quality assurance

The European Union Reference Laboratory for GM Food and Feed (EURL GMFF) hosted

by the Joint Research Centre (JRC-Ispra and JRC-Geel) is ISOIEC 170252005

accredited (certificate number Belac 268 TEST flexible scope for DNA extraction DNA

identification and real-time PCR) and ISOIEC 170432010 accredited (certificate

number Belac 268 PT proficiency test provider)

Confidentiality statement

The procedures used for the organisation of PTs are accredited according to ISO

170432010 and guarantee that the identity of the participants and the information

provided by them is treated as confidential The participants in this PT received a unique

laboratory code that is used throughout this report Laboratory codes of the National

Reference Laboratories appointed in line with Regulation (EC) No 8822004 will be

disclosed to DG SANTE for (long-term) performance assessment

List of abbreviations

CT Comparative test (here referred to as PT proficiency test)

EURL European Union Reference Laboratory

GMFF Genetically modified food and feed

kbp (kilo) thousand base pairs

LOQ Limit of quantification

mm Mass fraction or mass per mass percentage

NRL National Reference Laboratory

OCL Official Control Laboratory

qPCR Quantitative (real-time) Polymerase Chain Reaction

SD Standard deviation

σpt Standard deviation for proficiency assessment

Uu Expandedstandard measurement uncertainty

z and ζ Performance scores used to assess the measurement capability of a laboratory

EURL GMFF CT 0217 report

Contents

Executive summary 6

1 Introduction 7

2 Test items 8

21 Test item 1 8

22 Test item 2 10

3 Instructions to the participants 11

4 Results 12

41 Participation to the PT 12

42 Information on the testing provided in the questionnaire 14

43 GM event identification 15

44 GM event quantification 15

441 Number of participants reporting a quantitative result 15

442 Assigned values 16

443 Calculation of performance scores 17

444 Performance of the laboratories 18

5 Conclusions 23

Acknowledgements 24

References 26

Annexes 28

Annex 1 Homogeneity and stability of test items 28

A11 Homogeneity of test items 28

A12 Stability of test items 29

Annex 2 Questionnaire data 30

Annex 3 Qualitative performance of the participants 38

Annex 4 Participants quantitative performance 39

EURL GMFF CT 0217 report

646

Executive summary

The European Union Reference Laboratory for Genetically Modified Food and Feed (EURL

GMFF) organised a proficiency test (PT) for National Reference Laboratories (NRLs) to

support the official controls on food and feed in line with Regulation (EC) No 8822004

Other official control laboratories were allowed to participate on a voluntary basis

Two test items were distributed test item 1 (T1) was composed of ground chicken feed

spiked with a mixture of GM soybean event MON-Oslash4Oslash32-6 (40-3-2) and non-GM soybean

and test item 2 (T2) was a soybean flour containing the same GM soybean event 40-3-2

Participants were required to screen T1 and T2 for the presence of three GM soybean

events and to quantify the event that was present with the highest GM mass fraction The

results had to be reported in GM mass fraction (massmass )

Eighty-six participants from 39 countries participated to this PT including 54 NRLs of

which 33 are designated in line with Regulation (EC) No 8822004 (NRL882) and 21 are

nominated in Regulation (EU) No 1202014 to support the EURL GMFF on method

validation (NRL120) as well as 10 other EU official control laboratories

The qualitative results ie the correct identification of the GM event were evaluated and

scored as correct or incorrect The assigned value for the 40-3-2 soybean mass fraction in

both test materials was derived as the robust mean of the data provided by the NRLs

Laboratory performance was primarily evaluated by calculating z scores

The results reported indicate that all participants identified the correct GM event in both

test items All but one NRL obtained a satisfactory performance (z) score for the reported

40-3-2 soybean mass fraction in both test items and the performance of one other NRL

was unacceptable for T1 because they reported the 40-3-2 mass fraction as below their

LOQ Six and two NRLs obtained a questionable z score for T1 and T2 respectively

Considering the results provided by the other participants (non-NRL) three and two non-

NRLs obtained an unsatisfactory z score for T1 and T2 respectively Another three

laboratories obtained a questionable z score for T1 and one laboratory for T2

The laboratories ability to provide results close to the assigned value within their claimed

measurement uncertainty was additionally evaluated by ζ scores Twenty four and 16

laboratories had an unsatisfactory ζ scores for T1 and T2 respectively Unsatisfactory ζ

scores were mainly the consequence of an underestimated or not reported measurement

uncertainty Guidance is provided for correctly estimating the measurement uncertainty of

analytical results

A root-cause analysis will be requested from NRLs with an unsatisfactory z score result in

this PT and will be followed-up

EURL GMFF CT 0217 report

746

1 Introduction

The Joint Research Centre (JRC) of the European Commission was established as European

Union Reference Laboratory for GM Food and Feed (EURL GMFF) by Regulations (EC) No

18292003(1) and (EC) No 8822004(2) Regulation (EC) No 8822004 also requires

Member States to designate National Reference Laboratories (NRL882) for each EURL

coordinating activities for the official control of compliance with food and feed law The

analytical methods used for these controls have been validated by the EURL GMFF as

required by Regulation (EC) No 18292003 and for this task the EURL GMFF is supported

by NRLs listed in Regulation (EU) No 1202014(3) (NRL120 several of them are also

NRL882) The Member States of the European Union may also appoint other laboratories

(non-NRLs) to perform the official controls on food and feed

It is crucial that official control laboratories can accurately and reliably determine the GM

content in food and feed samples Regulation (EC) No 18292003 established a threshold

for labelling of food and feed products containing genetically modified material that is

authorised in the EU (09 ) Furthermore Regulation (EU) No 6192011(4) introduced a

minimum performance limit (01 mm ) for detecting the accidental presence in feed of

genetically modified material with pending or expired authorisation status Compliance

with these values is verified by the Member States of the European Union in the official

control of food and feed

The EURL GMFF is tasked with the organisation of proficiency tests (called comparative

tests or CT in the GMO legislation(2)) to foster the correct application of the analytical

methods available for the official controls The EURL GMFF is operating under a quality

management system which is accredited according to ISOIEC 17043(5) for the

organisation of proficiency testing

This report summarises the results obtained in a PT organised by the EURL GMFF in 2017

(CT 0217) Participation in such PTs is mandatory for NRL882 recommended for

NRL120 and open to other official control laboratories

EURL GMFF CT 0217 report

846

2 Test items

The test items used in this PT were prepared and characterised at JRC-Geel

21 Test item 1

The T1 test item was prepared from base materials that were characterised before their

use (Table 1) The base materials employed for the preparation of T1 were chicken feed

(AVEVE for biological agriculture according to EC 8342007 and EC 8892008) the ERM-

BF410ak as non GM soybean powder and the ERM-BF410bp(6) containing the MON-Oslash4Oslash32-

6 event (hereafter named 40-3-2) as spiking material (Table 1)

The chicken feed was composed according to the label of Bio maize kernels Bio soybean

oil-cake Bio-sunflower oil-cake Bio wheat Bio barley maize gluten potato proteins

calcium monophosphate calcium carbonate Bio soybean oil and sodium chloride The

analytical composition indicated a content of 175 protein 5 fat 13 ashes 5

cellulose 033 methionine 078 lysine 37 calcium 052 phosphorus and

01 sodium The chicken feed was milled using a cryo-grinding vibrating mill (Palla mill

KHD Humboldt-Wedag Koumlln DE) and sieved with a 500 microm stainless steel mesh on a

sieving machine equipped with an ultrasonic sieving aid (Russel Finex London UK) The

remaining powder was mixed in a DynaMIX CM200 (WAB Muttenz CH) for 1 h to

homogenise the distribution of the different types of seed tissues

The residual water mass fractions for the chicken feed powder and the powders of the

certified reference materials ERM-BF410ak and ERM-BF410bp were measured by

volumetric Karl Fischer titration (758 KFD Titrino Metrohm Herisau CH) The results

showed that the powders were sufficiently dry to perform the dry mixing and did not

require an additional drying step

The particle size distribution of the powders was measured using laser diffraction (PSA

Sympatec Clausthal-Zellerfeld DE) It was concluded that the particle size distribution of

these powders was sufficiently similar to allow subsequent preparations of mixtures

The amount and the quality of the DNA extracted from the chicken feed powder the non-

GM soybean flour and the GM spiking material were verified by UV spectrometry

fluorometry and gel electrophoresis A CTAB-tip 20G method (Qiagen Hilden Germany)

optimised for soybean was chosen with a sample intake of 200 mg because it yielded a

sufficient amount of DNA of PCR-grade quality from the base materials DNA extracted

with the in-house CTAB method was tested for PCR inhibition between 40 ngmicroL to 02

ngmicroL with a lectin qPCR assay (5 microL per PCR) and did not show any inhibition (Cq values

were very close to the theoretical Cq values) The PCR efficiencies ranged from 96 to 98

with a coefficient of determination (R2) between 099 and 100 confirming the absence

of significant amounts of PCR inhibitors in the extracts

The CTAB method yielded a sufficient amount of DNA of PCR-grade quality from both non-

GM and GM base materials

The level of fragmentation of the extracted DNA was investigated by 10 agarose gel

electrophoresis A smear from plusmn 12 to 1 kbp could be clearly seen in the chicken feed

DNA indicating some level of fragmentation of the extracted DNA while the DNA

extracted from the soybean materials migrated as a high molecular weight band (above 12

kbp) The amount of soybean DNA that could be extracted and amplified from the chicken

feed powder was determined by qPCR with a lectin assay using DNA from a soybean CRM

as calibrant this amount appeared to be rather low (lt1 of total DNA) The yield of

amplifiable DNA per mg of chicken feed powder and the yield of DNA measured by

PicoGreen for the soybean materials composed of pure soybean were taken into account

to calculate the amount of chicken feed powder non-GM soybean and GM soybean to be

mixed to obtain a target value of approximately 08 mm event 40-3-2 in T1

EURL GMFF CT 0217 report

946

Table 1 Characteristics of the base materials used for preparation of test item 1 (T1)

Characteristic Chicken Feed Non-GM

Soybean Flour

40-3-2

Soybean Flour

Type of base material Scratch grains CRM CRM

Origin AVEVE (Belgium) ERM-BF410ak ERM-BF410bp

Grinding method Cryo-grinding vibrating mill Used as such Used as such

Mixing method DynaMIX CM200 (WAB Muttenz CH)

Water content in gkg mean plusmn U

(k = 2 n = 3) 106 plusmn 15 115 plusmn 07 155 plusmn 10

Particle diameter in microm mean plusmn U

(k = 2 n = 5) 1114 plusmn 198 1188 plusmn 272 1213 plusmn 60

Soybean DNA yield in ngmg1 mean plusmn U

(k = 2) 07 plusmn 01 (n=5) 579 plusmn 178 (n=3) 729 plusmn 18 (n=2)

Genetic elements detected with screening

pre-spotted plates (Cq value)2

Hmg (Cq 246) Lec (247) CruA (Cq 357) UGP (281) PLD (365) P35S (Cq 333) tNOS (Cq 368) CTP2-EPSPS (Cq 357) Cry1AbAc (Cq 360) PAT (Cq 376)

Lec (Cq 215) Lec (Cq 212) p35S (Cq 212) tNOS (Cq 225) PAT (Cq 348)

GM soybean events detected with event-

specific pre-spotted plates (Cq value)2

40-3-2 (Cq 340) MON87701 (Cq 374) MON89788 (Cq 363) Lec (Cq 246)

NA 40-3-2 (Cq 214) A5547 (Cq 412)

Mass used to prepare T1 (g) 11032 96 108

Nominal target GM mass fraction in T1

(mm ) NA NA 08

1 Results reported here for a sample intake of 200 mg with the in-house validated CTAB method + Genomic-tip 20G purification for soybean (JRC-GEEL) The soybean DNA yield value for the chicken feed was determined by qPCR whereas the yield from the non-GM and GM soybean materials was measured by fluorometry 2 A screening and GM soybean event-specific pre-spotted plate (PSP) was used for these tests NA not applicable k coverage factor U expanded uncertainty

The presence of different species and GM events in the base materials and in a pilot

mixture was tested by using the screening(7) and GM soybean event-specific pre-spotted

plates(8)

The presence of maize and soybean in the chicken feed powder was confirmed by the early

quantification cycle obtained for the high mobility gene (hmg) and lectin (lec) assays Late

amplifications for the UDP-glucose pyrophosphorylase (ugp) phospholipase D (PLD) and

cruciferin A (cruA) genes confirmed the respective presence of potato sunflower and

rapeseed in the chicken feed

The chicken feed powder (labelled bio) also contained traces of genetic markers such as

p35S tNOS EPSPS and Cry1AbAc which indicate a contamination of the chicken feed by

genetically modified plant materials The GM soybean event-specific assays confirmed the

presence of traces of 3 GM soybeans namely 40-3-2 MON87701 and MON89788 The

level of contamination was estimated to be below 001

The final test item was prepared gravimetrically in accordance with ISO 170342016(9) as

follows

The mass of the GM ingredient to add (40-3-2 soybean) was calculated taking into

account the amount of DNA that could be extracted and amplified from the different

materials (Table 1)

The compound sample T1 was mixed in a DynaMIX CM200 for 15 h to improve

homogeneity

After finalisation of the mixing step the powders were filled manually in 20 mL

brown glass vials using lyophilisation inserts manually placed in the bottle necks

Before final closure of the vials air was evacuated in a freeze-dryer and replaced

by argon The vials were finally closed inside the freeze-dryer with the help of a

hydraulic device and then sealed with blue aluminium caps to maintain the inert

atmosphere and to prevent accidental opening during storage and transport

A total of 200 vials containing each at least 5 g of flour were then labelled with a

sample number and the description Sample T1 (chicken feed)

EURL GMFF CT 0217 report

1046

Following the inventory and the selection of vials for future analysis according to a

random stratified sampling scheme the bottles were brought to a storage room for

long-term storage in the dark at 4 plusmn 3 degC

Homogeneity and stability testing of T1 was performed in-house as described in Annex 1

using an event-specific quantification method previously validated by the EURL GMFF

Material T1 was found to be homogeneous for the GM event added (p-value gt 005)

based on a 200 mg sample intake

From the isochronous stability study it was concluded that the test item would be

sufficiently stable under ambient shipment conditions (5 significance level) Stability

was also confirmed during the whole period of the PT between the dispatch of the test

items until the deadline of reporting the results (Annex 1)

JRC-Ispra tested the T1 material and this confirmed the results obtained by JRC-Geel The

average (n = 91) mass fraction of event 40-3-2 measured in T1 was 069 plusmn 004 mm

(U k = 2) which approximated the expected nominal value but may have been

influenced by the characteristics of the different base materials

22 Test item 2

The T2 test item was a new batch of a certified reference material that was not yet

released on the market (Table 2) The bottles of T2 were re-labelled with a unique sample

number as Sample T2 (soybean flour)

Homogeneity and short-term stability of T2 had been previously demonstrated as part of

the certification of the CRM stability monitoring confirmed the stability of T2 during the

running time of the PT (Annex 1)

Table 2 Characteristics of test item 2 (T2)

Characteristic Soybean feed

Type of base material CRM

Origin ERM-BF410dp(6) containing 100 plusmn 06 gkg MON-Oslash4Oslash32-6 soybean produced in 2017 by JRC-Geel

The certificate of ERM-BF410dp warns that a difference (at 95 confidence level)

between the total DNA content in the two powders used for the production of ERM-

BF410dp was found to be significant (due to the different size of non-GM and GM seeds)

and is likely to have an impact when using this CRM Depending on the composition of the

unknown sample real-time PCR measurement results of ERM-BF410dp may differ up to

239 plusmn 11 (average plusmn U) compared to the results of the unknown sample This

difference may depend also on the DNA extraction method selected and both effects may

be additive This observation was confirmed by the participants in this PT who used the

previous 40-3-2 soybean CRM batch (ERM-BF410n) for the calibration of the T2

measurements

EURL GMFF CT 0217 report

1146

3 Instructions to the participants

Participants in this PT were instructed to analyse the two test items (T1 and T2) as

follows

Test Item 1 Chicken feed

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Test Item 2 Soybean flour

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Quantitative results had to be reported in mm as outlined below

Mass GM event [g]

mm = x 100 (1)

Total mass species [g]

Participants were requested

- to use procedures for detectionquantification of the GM events that resemble as

closely as possible the ones used in routine sample analysis

- to take care in ensuring that the DNA extraction method employed is adapted to

the matrix and that the quality of the DNA obtained is suitable for PCR

- to report the quantitative results with two decimal places (eg 064 or 129)

- to follow the general rule that results obtained using a calibrant certified for GM

mass fraction (ie a matrix CRM certified in [x] gkg) can directly be expressed in

mm while results obtained using a calibrant certified for DNA copy number

ratio (eg a plasmid containing both the GM and reference gene target or some

matrix CRMs) need to be converted into mm using a conversion factor(1011)

- to pay attention to the correct estimation and reporting of the measurement

uncertainty and coverage factor used as the uncertainty reported would be

considered in the evaluation of the results using zeta scores

- in case of an unsatisfactory performance to fill in a form indicating the root-cause

analysis and providing evidence demonstrating the effectiveness of the corrective

actions implemented (for NRLs only)

The participants were also informed that the identification information on the participants

in this PT would be kept confidential except for the NRLs that have been appointed in line

with Regulation (EC) No 8822004 their lab codes will be disclosed to DG SANTE for the

purpose of an assessment of their performance

EURL GMFF CT 0217 report

1246

4 Results

41 Participation to the PT

On 31 May 2017 199 laboratories were informed about the upcoming PT EURL-GMFF-CT-

0217 Finally 89 laboratories registered for it and received a random unique lab code

(L01 to L89) Eighty-six laboratories from 39 countries returned results within the

reporting deadline Three non-NRLs did not submit any results two of which (L21 L28)

had not received the samples from customs while L67 did not provide any justification for

not participating

Table 3 shows an overview of the participation to this PT

Table 3 Communication about and participation to the PT 0217

Characteristic of the PT Result

Date of PT announcement 31 May 2017

Deadline for registration 14 June 2017

Date of shipment of samples 3 July 2017

Deadline for result submission 25 August 2017

Number of laboratories informed 199

Number of registered laboratories 89

Registered laboratories that failed to submit their data 3

Number of participating laboratories 86

The participating laboratories fell into the following assigned categories (Table 4)

Thirty-three NRLs designated in line with Regulation (EC) No 8822004 (NRL882)

representing 25 EU Member States (many of them are also NRL120) In addition

Ireland delegated its NRL882 tasks to one of the PT participants Estonia and

Malta were not represented in this PT

Twenty-one NRLs nominated under Regulation (EU) No 1202014 (NRL120) that

are not at the same time NRL under Regulation (EC) No 8822004

Thirty-two laboratories that are not NRL but are appointed by their National

Authority to perform official controls This category includes 10 EU official control

laboratories (OCLs) and 22 laboratories from non-EU countries including Serbia

and Switzerland

Among the countries Germany was represented with 17 laboratories Italy with 6

laboratories and Belgium and Poland with 4 laboratories each all other countries had

between one and three participating laboratories

EURL GMFF CT 0217 report

1346

Table 4 Overview of participants to CT 0217 by country and category

Country Participants NRL8821 NRL120 Non-NRL

AUSTRIA 2 2

BELGIUM 4 3

1

BRAZIL 2

2

BULGARIA 2 1

1

CHILE 1

1

COLOMBIA 1

1

CROATIA 2 1

1

CYPRUS 1 1

CZECH REPUBLIC 1 1

DENMARK 1 1

FINLAND 2 1 1

FRANCE 2 2

GERMANY 17 1 14 2

GREECE 1 1

HUNGARY 2 1

1

INDIA 1

1

ITALY 6 1 2 3

LATVIA 1 1

LEBANON 1

1

LITHUANIA 1 1

LUXEMBOURG 1 1

MEXICO 1

1

NETHERLANDS 2 1 1

PHILIPPINES 1

1

POLAND 4 3 1

PORTUGAL 1 1

ROMANIA 2 1

1

SERBIA 3

3

SINGAPORE 1

1

SLOVAKIA 2 2

SLOVENIA 1 1

SPAIN 2 2

SWEDEN 1 1

SWITZERLAND 2

2

TURKEY 1

1

UKRAINE 3

3

UNITED KINGDOM 3 1 2

UNITED STATES 1

1

VIETNAM 3

3

Total 86 33 21 32 1 No NRL882 from Estonia or Malta participated to this PT

EURL GMFF CT 0217 report

1446

42 Information on the testing provided in the questionnaire

Participants were asked to fill in an online questionnaire (through EUSurvey) on their

testing methodology used for T1 and T2 consisting of a number of mostly multiple-choice

questions A total of 82 laboratories completed the questionnaire including all 33

NRL882 20 out of 21 NRL120 and 29 out of 32 non-NRLs (questionnaires were missing

from L03 L10 L38 and L49)

Table 5 summarises the main answers received whereas Annex 2 shows all answers The

results on GM event identification are reported in Section 43

Table 5 Summary of the main answers provided in the questionnaire of CT 0217

Question (and Question

number) Test Item 1 ndash 40-3-2 Test Item 2 ndash 40-3-2

Test item analysed Yes (801) No (2) Yes (82) No (0)

Reason for lack of analysis (Q1)

Matrix out of scope (1) other practical constraints (1) -

DNA extraction method (Q2) CTAB (41) NucleoSpin Food (10) CTAB (40) NucleoSpin Food (11)

Additional DNA purification method (Q3)

None (51) Ethanol (9) None (52) Ethanol (9)

Number of replicates (Q4) 2 (52) 4 (10) 2 (56) 4 (10)

Approach to test for PCR inhibition (Q5)

OD ratios (40) delta Cq or GM between two dilutions (31)

OD ratios (36) delta Cq or GM between two dilutions (32)

Reason for not testing all events (Q9)

Not applicable (45) below the LOQ (18) Not applicable (61) reagents not available (10)

Approach used (Q6a) Standard curves (68) delta Cq (9) Standard curves (71) delta Cq (9)

Calibrant used (Q6b) CRM JRC-Geel in gkg (71) other RM in copies (4) CRM JRC-Geel in gkg (74) other RM in copies (4)

Taxon-specific endogenous gene (Q6c)

lec-74 bp (65) lec-81 or 118 bp (each 5) lec-74 bp (63) lec-118 bp (6)

Unit of measurement and data expression (Q6d)

Mass (59) copies=mass CRM (13) Mass (61) copies=mass CRM (14)

Amount of DNA (Q6e) 200 ng (30) 100 ng (18) 200 ng (30) 100 ng (21)

LOQ (Q6f) 01 (37) lt01 (34) 01 (39) lt01 (39)

LOQ determination (Q6g) In-house validation (42) current analysis (20) In-house validation (43) current analysis (21)

Uncertainty determination (Q6h)

Precision of replicates (32) in-house validation (27) Precision of replicates (32) in-house validation (30)

1 The numbers shown refer to the number of laboratories that reported the answer The answers that were reported with the two largest frequencies are mentioned

One NRL882 (L73) reported that T1 was out of the scope of the laboratory and one non-

NRL (L43) reported that the T1 matrix was not analysed because of practical constraints

The evaluation of the answers shows that the most commonly employed DNA extraction

method for both T1 and T2 was one based on CTAB with the NucleoSpin Food kit ranking

second No additional purification methods were generally applied The majority of

laboratories analysed two replicate DNA extracts Most laboratories checked the quality of

the DNA extracts by verifying the OD ratios andor running two dilutions a minority of

laboratories performed a PCR inhibition run on 3 or 4 DNA dilutions with a reference gene

For the quantitative analysis the most common approach used was based on two standard

curves however 9 laboratories applied the delta Cq approach One laboratory (L50)

mentioned the use of digital PCR for 40-3-2 soybean quantification in T1 and T2 The

available CRMs from the JRC were used by most laboratories but 5 laboratories used a

non-certified reference material (RM) where values were expressed in GM copy number

ratio (4) or GM mass fraction (1) Lec was used as taxon-specific reference gene by all

laboratories for soybean (mostly the 74 bp version) The majority of laboratories

performed their measurements in the same unit as the certified value of the calibrant used

(gkg) and no conversion factor was applied The LOQ reported was either taken from in-

house validation of the method or determined from the analysis results for this CT In

most cases a LOQ of 01 mm or lower was reported The measurement uncertainty

was either derived from the standard deviation of the measurement replicates or from the

intermediate precision determined in the frame of the single-laboratory validation study

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

Quality assurance

The European Union Reference Laboratory for GM Food and Feed (EURL GMFF) hosted

by the Joint Research Centre (JRC-Ispra and JRC-Geel) is ISOIEC 170252005

accredited (certificate number Belac 268 TEST flexible scope for DNA extraction DNA

identification and real-time PCR) and ISOIEC 170432010 accredited (certificate

number Belac 268 PT proficiency test provider)

Confidentiality statement

The procedures used for the organisation of PTs are accredited according to ISO

170432010 and guarantee that the identity of the participants and the information

provided by them is treated as confidential The participants in this PT received a unique

laboratory code that is used throughout this report Laboratory codes of the National

Reference Laboratories appointed in line with Regulation (EC) No 8822004 will be

disclosed to DG SANTE for (long-term) performance assessment

List of abbreviations

CT Comparative test (here referred to as PT proficiency test)

EURL European Union Reference Laboratory

GMFF Genetically modified food and feed

kbp (kilo) thousand base pairs

LOQ Limit of quantification

mm Mass fraction or mass per mass percentage

NRL National Reference Laboratory

OCL Official Control Laboratory

qPCR Quantitative (real-time) Polymerase Chain Reaction

SD Standard deviation

σpt Standard deviation for proficiency assessment

Uu Expandedstandard measurement uncertainty

z and ζ Performance scores used to assess the measurement capability of a laboratory

EURL GMFF CT 0217 report

Contents

Executive summary 6

1 Introduction 7

2 Test items 8

21 Test item 1 8

22 Test item 2 10

3 Instructions to the participants 11

4 Results 12

41 Participation to the PT 12

42 Information on the testing provided in the questionnaire 14

43 GM event identification 15

44 GM event quantification 15

441 Number of participants reporting a quantitative result 15

442 Assigned values 16

443 Calculation of performance scores 17

444 Performance of the laboratories 18

5 Conclusions 23

Acknowledgements 24

References 26

Annexes 28

Annex 1 Homogeneity and stability of test items 28

A11 Homogeneity of test items 28

A12 Stability of test items 29

Annex 2 Questionnaire data 30

Annex 3 Qualitative performance of the participants 38

Annex 4 Participants quantitative performance 39

EURL GMFF CT 0217 report

646

Executive summary

The European Union Reference Laboratory for Genetically Modified Food and Feed (EURL

GMFF) organised a proficiency test (PT) for National Reference Laboratories (NRLs) to

support the official controls on food and feed in line with Regulation (EC) No 8822004

Other official control laboratories were allowed to participate on a voluntary basis

Two test items were distributed test item 1 (T1) was composed of ground chicken feed

spiked with a mixture of GM soybean event MON-Oslash4Oslash32-6 (40-3-2) and non-GM soybean

and test item 2 (T2) was a soybean flour containing the same GM soybean event 40-3-2

Participants were required to screen T1 and T2 for the presence of three GM soybean

events and to quantify the event that was present with the highest GM mass fraction The

results had to be reported in GM mass fraction (massmass )

Eighty-six participants from 39 countries participated to this PT including 54 NRLs of

which 33 are designated in line with Regulation (EC) No 8822004 (NRL882) and 21 are

nominated in Regulation (EU) No 1202014 to support the EURL GMFF on method

validation (NRL120) as well as 10 other EU official control laboratories

The qualitative results ie the correct identification of the GM event were evaluated and

scored as correct or incorrect The assigned value for the 40-3-2 soybean mass fraction in

both test materials was derived as the robust mean of the data provided by the NRLs

Laboratory performance was primarily evaluated by calculating z scores

The results reported indicate that all participants identified the correct GM event in both

test items All but one NRL obtained a satisfactory performance (z) score for the reported

40-3-2 soybean mass fraction in both test items and the performance of one other NRL

was unacceptable for T1 because they reported the 40-3-2 mass fraction as below their

LOQ Six and two NRLs obtained a questionable z score for T1 and T2 respectively

Considering the results provided by the other participants (non-NRL) three and two non-

NRLs obtained an unsatisfactory z score for T1 and T2 respectively Another three

laboratories obtained a questionable z score for T1 and one laboratory for T2

The laboratories ability to provide results close to the assigned value within their claimed

measurement uncertainty was additionally evaluated by ζ scores Twenty four and 16

laboratories had an unsatisfactory ζ scores for T1 and T2 respectively Unsatisfactory ζ

scores were mainly the consequence of an underestimated or not reported measurement

uncertainty Guidance is provided for correctly estimating the measurement uncertainty of

analytical results

A root-cause analysis will be requested from NRLs with an unsatisfactory z score result in

this PT and will be followed-up

EURL GMFF CT 0217 report

746

1 Introduction

The Joint Research Centre (JRC) of the European Commission was established as European

Union Reference Laboratory for GM Food and Feed (EURL GMFF) by Regulations (EC) No

18292003(1) and (EC) No 8822004(2) Regulation (EC) No 8822004 also requires

Member States to designate National Reference Laboratories (NRL882) for each EURL

coordinating activities for the official control of compliance with food and feed law The

analytical methods used for these controls have been validated by the EURL GMFF as

required by Regulation (EC) No 18292003 and for this task the EURL GMFF is supported

by NRLs listed in Regulation (EU) No 1202014(3) (NRL120 several of them are also

NRL882) The Member States of the European Union may also appoint other laboratories

(non-NRLs) to perform the official controls on food and feed

It is crucial that official control laboratories can accurately and reliably determine the GM

content in food and feed samples Regulation (EC) No 18292003 established a threshold

for labelling of food and feed products containing genetically modified material that is

authorised in the EU (09 ) Furthermore Regulation (EU) No 6192011(4) introduced a

minimum performance limit (01 mm ) for detecting the accidental presence in feed of

genetically modified material with pending or expired authorisation status Compliance

with these values is verified by the Member States of the European Union in the official

control of food and feed

The EURL GMFF is tasked with the organisation of proficiency tests (called comparative

tests or CT in the GMO legislation(2)) to foster the correct application of the analytical

methods available for the official controls The EURL GMFF is operating under a quality

management system which is accredited according to ISOIEC 17043(5) for the

organisation of proficiency testing

This report summarises the results obtained in a PT organised by the EURL GMFF in 2017

(CT 0217) Participation in such PTs is mandatory for NRL882 recommended for

NRL120 and open to other official control laboratories

EURL GMFF CT 0217 report

846

2 Test items

The test items used in this PT were prepared and characterised at JRC-Geel

21 Test item 1

The T1 test item was prepared from base materials that were characterised before their

use (Table 1) The base materials employed for the preparation of T1 were chicken feed

(AVEVE for biological agriculture according to EC 8342007 and EC 8892008) the ERM-

BF410ak as non GM soybean powder and the ERM-BF410bp(6) containing the MON-Oslash4Oslash32-

6 event (hereafter named 40-3-2) as spiking material (Table 1)

The chicken feed was composed according to the label of Bio maize kernels Bio soybean

oil-cake Bio-sunflower oil-cake Bio wheat Bio barley maize gluten potato proteins

calcium monophosphate calcium carbonate Bio soybean oil and sodium chloride The

analytical composition indicated a content of 175 protein 5 fat 13 ashes 5

cellulose 033 methionine 078 lysine 37 calcium 052 phosphorus and

01 sodium The chicken feed was milled using a cryo-grinding vibrating mill (Palla mill

KHD Humboldt-Wedag Koumlln DE) and sieved with a 500 microm stainless steel mesh on a

sieving machine equipped with an ultrasonic sieving aid (Russel Finex London UK) The

remaining powder was mixed in a DynaMIX CM200 (WAB Muttenz CH) for 1 h to

homogenise the distribution of the different types of seed tissues

The residual water mass fractions for the chicken feed powder and the powders of the

certified reference materials ERM-BF410ak and ERM-BF410bp were measured by

volumetric Karl Fischer titration (758 KFD Titrino Metrohm Herisau CH) The results

showed that the powders were sufficiently dry to perform the dry mixing and did not

require an additional drying step

The particle size distribution of the powders was measured using laser diffraction (PSA

Sympatec Clausthal-Zellerfeld DE) It was concluded that the particle size distribution of

these powders was sufficiently similar to allow subsequent preparations of mixtures

The amount and the quality of the DNA extracted from the chicken feed powder the non-

GM soybean flour and the GM spiking material were verified by UV spectrometry

fluorometry and gel electrophoresis A CTAB-tip 20G method (Qiagen Hilden Germany)

optimised for soybean was chosen with a sample intake of 200 mg because it yielded a

sufficient amount of DNA of PCR-grade quality from the base materials DNA extracted

with the in-house CTAB method was tested for PCR inhibition between 40 ngmicroL to 02

ngmicroL with a lectin qPCR assay (5 microL per PCR) and did not show any inhibition (Cq values

were very close to the theoretical Cq values) The PCR efficiencies ranged from 96 to 98

with a coefficient of determination (R2) between 099 and 100 confirming the absence

of significant amounts of PCR inhibitors in the extracts

The CTAB method yielded a sufficient amount of DNA of PCR-grade quality from both non-

GM and GM base materials

The level of fragmentation of the extracted DNA was investigated by 10 agarose gel

electrophoresis A smear from plusmn 12 to 1 kbp could be clearly seen in the chicken feed

DNA indicating some level of fragmentation of the extracted DNA while the DNA

extracted from the soybean materials migrated as a high molecular weight band (above 12

kbp) The amount of soybean DNA that could be extracted and amplified from the chicken

feed powder was determined by qPCR with a lectin assay using DNA from a soybean CRM

as calibrant this amount appeared to be rather low (lt1 of total DNA) The yield of

amplifiable DNA per mg of chicken feed powder and the yield of DNA measured by

PicoGreen for the soybean materials composed of pure soybean were taken into account

to calculate the amount of chicken feed powder non-GM soybean and GM soybean to be

mixed to obtain a target value of approximately 08 mm event 40-3-2 in T1

EURL GMFF CT 0217 report

946

Table 1 Characteristics of the base materials used for preparation of test item 1 (T1)

Characteristic Chicken Feed Non-GM

Soybean Flour

40-3-2

Soybean Flour

Type of base material Scratch grains CRM CRM

Origin AVEVE (Belgium) ERM-BF410ak ERM-BF410bp

Grinding method Cryo-grinding vibrating mill Used as such Used as such

Mixing method DynaMIX CM200 (WAB Muttenz CH)

Water content in gkg mean plusmn U

(k = 2 n = 3) 106 plusmn 15 115 plusmn 07 155 plusmn 10

Particle diameter in microm mean plusmn U

(k = 2 n = 5) 1114 plusmn 198 1188 plusmn 272 1213 plusmn 60

Soybean DNA yield in ngmg1 mean plusmn U

(k = 2) 07 plusmn 01 (n=5) 579 plusmn 178 (n=3) 729 plusmn 18 (n=2)

Genetic elements detected with screening

pre-spotted plates (Cq value)2

Hmg (Cq 246) Lec (247) CruA (Cq 357) UGP (281) PLD (365) P35S (Cq 333) tNOS (Cq 368) CTP2-EPSPS (Cq 357) Cry1AbAc (Cq 360) PAT (Cq 376)

Lec (Cq 215) Lec (Cq 212) p35S (Cq 212) tNOS (Cq 225) PAT (Cq 348)

GM soybean events detected with event-

specific pre-spotted plates (Cq value)2

40-3-2 (Cq 340) MON87701 (Cq 374) MON89788 (Cq 363) Lec (Cq 246)

NA 40-3-2 (Cq 214) A5547 (Cq 412)

Mass used to prepare T1 (g) 11032 96 108

Nominal target GM mass fraction in T1

(mm ) NA NA 08

1 Results reported here for a sample intake of 200 mg with the in-house validated CTAB method + Genomic-tip 20G purification for soybean (JRC-GEEL) The soybean DNA yield value for the chicken feed was determined by qPCR whereas the yield from the non-GM and GM soybean materials was measured by fluorometry 2 A screening and GM soybean event-specific pre-spotted plate (PSP) was used for these tests NA not applicable k coverage factor U expanded uncertainty

The presence of different species and GM events in the base materials and in a pilot

mixture was tested by using the screening(7) and GM soybean event-specific pre-spotted

plates(8)

The presence of maize and soybean in the chicken feed powder was confirmed by the early

quantification cycle obtained for the high mobility gene (hmg) and lectin (lec) assays Late

amplifications for the UDP-glucose pyrophosphorylase (ugp) phospholipase D (PLD) and

cruciferin A (cruA) genes confirmed the respective presence of potato sunflower and

rapeseed in the chicken feed

The chicken feed powder (labelled bio) also contained traces of genetic markers such as

p35S tNOS EPSPS and Cry1AbAc which indicate a contamination of the chicken feed by

genetically modified plant materials The GM soybean event-specific assays confirmed the

presence of traces of 3 GM soybeans namely 40-3-2 MON87701 and MON89788 The

level of contamination was estimated to be below 001

The final test item was prepared gravimetrically in accordance with ISO 170342016(9) as

follows

The mass of the GM ingredient to add (40-3-2 soybean) was calculated taking into

account the amount of DNA that could be extracted and amplified from the different

materials (Table 1)

The compound sample T1 was mixed in a DynaMIX CM200 for 15 h to improve

homogeneity

After finalisation of the mixing step the powders were filled manually in 20 mL

brown glass vials using lyophilisation inserts manually placed in the bottle necks

Before final closure of the vials air was evacuated in a freeze-dryer and replaced

by argon The vials were finally closed inside the freeze-dryer with the help of a

hydraulic device and then sealed with blue aluminium caps to maintain the inert

atmosphere and to prevent accidental opening during storage and transport

A total of 200 vials containing each at least 5 g of flour were then labelled with a

sample number and the description Sample T1 (chicken feed)

EURL GMFF CT 0217 report

1046

Following the inventory and the selection of vials for future analysis according to a

random stratified sampling scheme the bottles were brought to a storage room for

long-term storage in the dark at 4 plusmn 3 degC

Homogeneity and stability testing of T1 was performed in-house as described in Annex 1

using an event-specific quantification method previously validated by the EURL GMFF

Material T1 was found to be homogeneous for the GM event added (p-value gt 005)

based on a 200 mg sample intake

From the isochronous stability study it was concluded that the test item would be

sufficiently stable under ambient shipment conditions (5 significance level) Stability

was also confirmed during the whole period of the PT between the dispatch of the test

items until the deadline of reporting the results (Annex 1)

JRC-Ispra tested the T1 material and this confirmed the results obtained by JRC-Geel The

average (n = 91) mass fraction of event 40-3-2 measured in T1 was 069 plusmn 004 mm

(U k = 2) which approximated the expected nominal value but may have been

influenced by the characteristics of the different base materials

22 Test item 2

The T2 test item was a new batch of a certified reference material that was not yet

released on the market (Table 2) The bottles of T2 were re-labelled with a unique sample

number as Sample T2 (soybean flour)

Homogeneity and short-term stability of T2 had been previously demonstrated as part of

the certification of the CRM stability monitoring confirmed the stability of T2 during the

running time of the PT (Annex 1)

Table 2 Characteristics of test item 2 (T2)

Characteristic Soybean feed

Type of base material CRM

Origin ERM-BF410dp(6) containing 100 plusmn 06 gkg MON-Oslash4Oslash32-6 soybean produced in 2017 by JRC-Geel

The certificate of ERM-BF410dp warns that a difference (at 95 confidence level)

between the total DNA content in the two powders used for the production of ERM-

BF410dp was found to be significant (due to the different size of non-GM and GM seeds)

and is likely to have an impact when using this CRM Depending on the composition of the

unknown sample real-time PCR measurement results of ERM-BF410dp may differ up to

239 plusmn 11 (average plusmn U) compared to the results of the unknown sample This

difference may depend also on the DNA extraction method selected and both effects may

be additive This observation was confirmed by the participants in this PT who used the

previous 40-3-2 soybean CRM batch (ERM-BF410n) for the calibration of the T2

measurements

EURL GMFF CT 0217 report

1146

3 Instructions to the participants

Participants in this PT were instructed to analyse the two test items (T1 and T2) as

follows

Test Item 1 Chicken feed

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Test Item 2 Soybean flour

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Quantitative results had to be reported in mm as outlined below

Mass GM event [g]

mm = x 100 (1)

Total mass species [g]

Participants were requested

- to use procedures for detectionquantification of the GM events that resemble as

closely as possible the ones used in routine sample analysis

- to take care in ensuring that the DNA extraction method employed is adapted to

the matrix and that the quality of the DNA obtained is suitable for PCR

- to report the quantitative results with two decimal places (eg 064 or 129)

- to follow the general rule that results obtained using a calibrant certified for GM

mass fraction (ie a matrix CRM certified in [x] gkg) can directly be expressed in

mm while results obtained using a calibrant certified for DNA copy number

ratio (eg a plasmid containing both the GM and reference gene target or some

matrix CRMs) need to be converted into mm using a conversion factor(1011)

- to pay attention to the correct estimation and reporting of the measurement

uncertainty and coverage factor used as the uncertainty reported would be

considered in the evaluation of the results using zeta scores

- in case of an unsatisfactory performance to fill in a form indicating the root-cause

analysis and providing evidence demonstrating the effectiveness of the corrective

actions implemented (for NRLs only)

The participants were also informed that the identification information on the participants

in this PT would be kept confidential except for the NRLs that have been appointed in line

with Regulation (EC) No 8822004 their lab codes will be disclosed to DG SANTE for the

purpose of an assessment of their performance

EURL GMFF CT 0217 report

1246

4 Results

41 Participation to the PT

On 31 May 2017 199 laboratories were informed about the upcoming PT EURL-GMFF-CT-

0217 Finally 89 laboratories registered for it and received a random unique lab code

(L01 to L89) Eighty-six laboratories from 39 countries returned results within the

reporting deadline Three non-NRLs did not submit any results two of which (L21 L28)

had not received the samples from customs while L67 did not provide any justification for

not participating

Table 3 shows an overview of the participation to this PT

Table 3 Communication about and participation to the PT 0217

Characteristic of the PT Result

Date of PT announcement 31 May 2017

Deadline for registration 14 June 2017

Date of shipment of samples 3 July 2017

Deadline for result submission 25 August 2017

Number of laboratories informed 199

Number of registered laboratories 89

Registered laboratories that failed to submit their data 3

Number of participating laboratories 86

The participating laboratories fell into the following assigned categories (Table 4)

Thirty-three NRLs designated in line with Regulation (EC) No 8822004 (NRL882)

representing 25 EU Member States (many of them are also NRL120) In addition

Ireland delegated its NRL882 tasks to one of the PT participants Estonia and

Malta were not represented in this PT

Twenty-one NRLs nominated under Regulation (EU) No 1202014 (NRL120) that

are not at the same time NRL under Regulation (EC) No 8822004

Thirty-two laboratories that are not NRL but are appointed by their National

Authority to perform official controls This category includes 10 EU official control

laboratories (OCLs) and 22 laboratories from non-EU countries including Serbia

and Switzerland

Among the countries Germany was represented with 17 laboratories Italy with 6

laboratories and Belgium and Poland with 4 laboratories each all other countries had

between one and three participating laboratories

EURL GMFF CT 0217 report

1346

Table 4 Overview of participants to CT 0217 by country and category

Country Participants NRL8821 NRL120 Non-NRL

AUSTRIA 2 2

BELGIUM 4 3

1

BRAZIL 2

2

BULGARIA 2 1

1

CHILE 1

1

COLOMBIA 1

1

CROATIA 2 1

1

CYPRUS 1 1

CZECH REPUBLIC 1 1

DENMARK 1 1

FINLAND 2 1 1

FRANCE 2 2

GERMANY 17 1 14 2

GREECE 1 1

HUNGARY 2 1

1

INDIA 1

1

ITALY 6 1 2 3

LATVIA 1 1

LEBANON 1

1

LITHUANIA 1 1

LUXEMBOURG 1 1

MEXICO 1

1

NETHERLANDS 2 1 1

PHILIPPINES 1

1

POLAND 4 3 1

PORTUGAL 1 1

ROMANIA 2 1

1

SERBIA 3

3

SINGAPORE 1

1

SLOVAKIA 2 2

SLOVENIA 1 1

SPAIN 2 2

SWEDEN 1 1

SWITZERLAND 2

2

TURKEY 1

1

UKRAINE 3

3

UNITED KINGDOM 3 1 2

UNITED STATES 1

1

VIETNAM 3

3

Total 86 33 21 32 1 No NRL882 from Estonia or Malta participated to this PT

EURL GMFF CT 0217 report

1446

42 Information on the testing provided in the questionnaire

Participants were asked to fill in an online questionnaire (through EUSurvey) on their

testing methodology used for T1 and T2 consisting of a number of mostly multiple-choice

questions A total of 82 laboratories completed the questionnaire including all 33

NRL882 20 out of 21 NRL120 and 29 out of 32 non-NRLs (questionnaires were missing

from L03 L10 L38 and L49)

Table 5 summarises the main answers received whereas Annex 2 shows all answers The

results on GM event identification are reported in Section 43

Table 5 Summary of the main answers provided in the questionnaire of CT 0217

Question (and Question

number) Test Item 1 ndash 40-3-2 Test Item 2 ndash 40-3-2

Test item analysed Yes (801) No (2) Yes (82) No (0)

Reason for lack of analysis (Q1)

Matrix out of scope (1) other practical constraints (1) -

DNA extraction method (Q2) CTAB (41) NucleoSpin Food (10) CTAB (40) NucleoSpin Food (11)

Additional DNA purification method (Q3)

None (51) Ethanol (9) None (52) Ethanol (9)

Number of replicates (Q4) 2 (52) 4 (10) 2 (56) 4 (10)

Approach to test for PCR inhibition (Q5)

OD ratios (40) delta Cq or GM between two dilutions (31)

OD ratios (36) delta Cq or GM between two dilutions (32)

Reason for not testing all events (Q9)

Not applicable (45) below the LOQ (18) Not applicable (61) reagents not available (10)

Approach used (Q6a) Standard curves (68) delta Cq (9) Standard curves (71) delta Cq (9)

Calibrant used (Q6b) CRM JRC-Geel in gkg (71) other RM in copies (4) CRM JRC-Geel in gkg (74) other RM in copies (4)

Taxon-specific endogenous gene (Q6c)

lec-74 bp (65) lec-81 or 118 bp (each 5) lec-74 bp (63) lec-118 bp (6)

Unit of measurement and data expression (Q6d)

Mass (59) copies=mass CRM (13) Mass (61) copies=mass CRM (14)

Amount of DNA (Q6e) 200 ng (30) 100 ng (18) 200 ng (30) 100 ng (21)

LOQ (Q6f) 01 (37) lt01 (34) 01 (39) lt01 (39)

LOQ determination (Q6g) In-house validation (42) current analysis (20) In-house validation (43) current analysis (21)

Uncertainty determination (Q6h)

Precision of replicates (32) in-house validation (27) Precision of replicates (32) in-house validation (30)

1 The numbers shown refer to the number of laboratories that reported the answer The answers that were reported with the two largest frequencies are mentioned

One NRL882 (L73) reported that T1 was out of the scope of the laboratory and one non-

NRL (L43) reported that the T1 matrix was not analysed because of practical constraints

The evaluation of the answers shows that the most commonly employed DNA extraction

method for both T1 and T2 was one based on CTAB with the NucleoSpin Food kit ranking

second No additional purification methods were generally applied The majority of

laboratories analysed two replicate DNA extracts Most laboratories checked the quality of

the DNA extracts by verifying the OD ratios andor running two dilutions a minority of

laboratories performed a PCR inhibition run on 3 or 4 DNA dilutions with a reference gene

For the quantitative analysis the most common approach used was based on two standard

curves however 9 laboratories applied the delta Cq approach One laboratory (L50)

mentioned the use of digital PCR for 40-3-2 soybean quantification in T1 and T2 The

available CRMs from the JRC were used by most laboratories but 5 laboratories used a

non-certified reference material (RM) where values were expressed in GM copy number

ratio (4) or GM mass fraction (1) Lec was used as taxon-specific reference gene by all

laboratories for soybean (mostly the 74 bp version) The majority of laboratories

performed their measurements in the same unit as the certified value of the calibrant used

(gkg) and no conversion factor was applied The LOQ reported was either taken from in-

house validation of the method or determined from the analysis results for this CT In

most cases a LOQ of 01 mm or lower was reported The measurement uncertainty

was either derived from the standard deviation of the measurement replicates or from the

intermediate precision determined in the frame of the single-laboratory validation study

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

Contents

Executive summary 6

1 Introduction 7

2 Test items 8

21 Test item 1 8

22 Test item 2 10

3 Instructions to the participants 11

4 Results 12

41 Participation to the PT 12

42 Information on the testing provided in the questionnaire 14

43 GM event identification 15

44 GM event quantification 15

441 Number of participants reporting a quantitative result 15

442 Assigned values 16

443 Calculation of performance scores 17

444 Performance of the laboratories 18

5 Conclusions 23

Acknowledgements 24

References 26

Annexes 28

Annex 1 Homogeneity and stability of test items 28

A11 Homogeneity of test items 28

A12 Stability of test items 29

Annex 2 Questionnaire data 30

Annex 3 Qualitative performance of the participants 38

Annex 4 Participants quantitative performance 39

EURL GMFF CT 0217 report

646

Executive summary

The European Union Reference Laboratory for Genetically Modified Food and Feed (EURL

GMFF) organised a proficiency test (PT) for National Reference Laboratories (NRLs) to

support the official controls on food and feed in line with Regulation (EC) No 8822004

Other official control laboratories were allowed to participate on a voluntary basis

Two test items were distributed test item 1 (T1) was composed of ground chicken feed

spiked with a mixture of GM soybean event MON-Oslash4Oslash32-6 (40-3-2) and non-GM soybean

and test item 2 (T2) was a soybean flour containing the same GM soybean event 40-3-2

Participants were required to screen T1 and T2 for the presence of three GM soybean

events and to quantify the event that was present with the highest GM mass fraction The

results had to be reported in GM mass fraction (massmass )

Eighty-six participants from 39 countries participated to this PT including 54 NRLs of

which 33 are designated in line with Regulation (EC) No 8822004 (NRL882) and 21 are

nominated in Regulation (EU) No 1202014 to support the EURL GMFF on method

validation (NRL120) as well as 10 other EU official control laboratories

The qualitative results ie the correct identification of the GM event were evaluated and

scored as correct or incorrect The assigned value for the 40-3-2 soybean mass fraction in

both test materials was derived as the robust mean of the data provided by the NRLs

Laboratory performance was primarily evaluated by calculating z scores

The results reported indicate that all participants identified the correct GM event in both

test items All but one NRL obtained a satisfactory performance (z) score for the reported

40-3-2 soybean mass fraction in both test items and the performance of one other NRL

was unacceptable for T1 because they reported the 40-3-2 mass fraction as below their

LOQ Six and two NRLs obtained a questionable z score for T1 and T2 respectively

Considering the results provided by the other participants (non-NRL) three and two non-

NRLs obtained an unsatisfactory z score for T1 and T2 respectively Another three

laboratories obtained a questionable z score for T1 and one laboratory for T2

The laboratories ability to provide results close to the assigned value within their claimed

measurement uncertainty was additionally evaluated by ζ scores Twenty four and 16

laboratories had an unsatisfactory ζ scores for T1 and T2 respectively Unsatisfactory ζ

scores were mainly the consequence of an underestimated or not reported measurement

uncertainty Guidance is provided for correctly estimating the measurement uncertainty of

analytical results

A root-cause analysis will be requested from NRLs with an unsatisfactory z score result in

this PT and will be followed-up

EURL GMFF CT 0217 report

746

1 Introduction

The Joint Research Centre (JRC) of the European Commission was established as European

Union Reference Laboratory for GM Food and Feed (EURL GMFF) by Regulations (EC) No

18292003(1) and (EC) No 8822004(2) Regulation (EC) No 8822004 also requires

Member States to designate National Reference Laboratories (NRL882) for each EURL

coordinating activities for the official control of compliance with food and feed law The

analytical methods used for these controls have been validated by the EURL GMFF as

required by Regulation (EC) No 18292003 and for this task the EURL GMFF is supported

by NRLs listed in Regulation (EU) No 1202014(3) (NRL120 several of them are also

NRL882) The Member States of the European Union may also appoint other laboratories

(non-NRLs) to perform the official controls on food and feed

It is crucial that official control laboratories can accurately and reliably determine the GM

content in food and feed samples Regulation (EC) No 18292003 established a threshold

for labelling of food and feed products containing genetically modified material that is

authorised in the EU (09 ) Furthermore Regulation (EU) No 6192011(4) introduced a

minimum performance limit (01 mm ) for detecting the accidental presence in feed of

genetically modified material with pending or expired authorisation status Compliance

with these values is verified by the Member States of the European Union in the official

control of food and feed

The EURL GMFF is tasked with the organisation of proficiency tests (called comparative

tests or CT in the GMO legislation(2)) to foster the correct application of the analytical

methods available for the official controls The EURL GMFF is operating under a quality

management system which is accredited according to ISOIEC 17043(5) for the

organisation of proficiency testing

This report summarises the results obtained in a PT organised by the EURL GMFF in 2017

(CT 0217) Participation in such PTs is mandatory for NRL882 recommended for

NRL120 and open to other official control laboratories

EURL GMFF CT 0217 report

846

2 Test items

The test items used in this PT were prepared and characterised at JRC-Geel

21 Test item 1

The T1 test item was prepared from base materials that were characterised before their

use (Table 1) The base materials employed for the preparation of T1 were chicken feed

(AVEVE for biological agriculture according to EC 8342007 and EC 8892008) the ERM-

BF410ak as non GM soybean powder and the ERM-BF410bp(6) containing the MON-Oslash4Oslash32-

6 event (hereafter named 40-3-2) as spiking material (Table 1)

The chicken feed was composed according to the label of Bio maize kernels Bio soybean

oil-cake Bio-sunflower oil-cake Bio wheat Bio barley maize gluten potato proteins

calcium monophosphate calcium carbonate Bio soybean oil and sodium chloride The

analytical composition indicated a content of 175 protein 5 fat 13 ashes 5

cellulose 033 methionine 078 lysine 37 calcium 052 phosphorus and

01 sodium The chicken feed was milled using a cryo-grinding vibrating mill (Palla mill

KHD Humboldt-Wedag Koumlln DE) and sieved with a 500 microm stainless steel mesh on a

sieving machine equipped with an ultrasonic sieving aid (Russel Finex London UK) The

remaining powder was mixed in a DynaMIX CM200 (WAB Muttenz CH) for 1 h to

homogenise the distribution of the different types of seed tissues

The residual water mass fractions for the chicken feed powder and the powders of the

certified reference materials ERM-BF410ak and ERM-BF410bp were measured by

volumetric Karl Fischer titration (758 KFD Titrino Metrohm Herisau CH) The results

showed that the powders were sufficiently dry to perform the dry mixing and did not

require an additional drying step

The particle size distribution of the powders was measured using laser diffraction (PSA

Sympatec Clausthal-Zellerfeld DE) It was concluded that the particle size distribution of

these powders was sufficiently similar to allow subsequent preparations of mixtures

The amount and the quality of the DNA extracted from the chicken feed powder the non-

GM soybean flour and the GM spiking material were verified by UV spectrometry

fluorometry and gel electrophoresis A CTAB-tip 20G method (Qiagen Hilden Germany)

optimised for soybean was chosen with a sample intake of 200 mg because it yielded a

sufficient amount of DNA of PCR-grade quality from the base materials DNA extracted

with the in-house CTAB method was tested for PCR inhibition between 40 ngmicroL to 02

ngmicroL with a lectin qPCR assay (5 microL per PCR) and did not show any inhibition (Cq values

were very close to the theoretical Cq values) The PCR efficiencies ranged from 96 to 98

with a coefficient of determination (R2) between 099 and 100 confirming the absence

of significant amounts of PCR inhibitors in the extracts

The CTAB method yielded a sufficient amount of DNA of PCR-grade quality from both non-

GM and GM base materials

The level of fragmentation of the extracted DNA was investigated by 10 agarose gel

electrophoresis A smear from plusmn 12 to 1 kbp could be clearly seen in the chicken feed

DNA indicating some level of fragmentation of the extracted DNA while the DNA

extracted from the soybean materials migrated as a high molecular weight band (above 12

kbp) The amount of soybean DNA that could be extracted and amplified from the chicken

feed powder was determined by qPCR with a lectin assay using DNA from a soybean CRM

as calibrant this amount appeared to be rather low (lt1 of total DNA) The yield of

amplifiable DNA per mg of chicken feed powder and the yield of DNA measured by

PicoGreen for the soybean materials composed of pure soybean were taken into account

to calculate the amount of chicken feed powder non-GM soybean and GM soybean to be

mixed to obtain a target value of approximately 08 mm event 40-3-2 in T1

EURL GMFF CT 0217 report

946

Table 1 Characteristics of the base materials used for preparation of test item 1 (T1)

Characteristic Chicken Feed Non-GM

Soybean Flour

40-3-2

Soybean Flour

Type of base material Scratch grains CRM CRM

Origin AVEVE (Belgium) ERM-BF410ak ERM-BF410bp

Grinding method Cryo-grinding vibrating mill Used as such Used as such

Mixing method DynaMIX CM200 (WAB Muttenz CH)

Water content in gkg mean plusmn U

(k = 2 n = 3) 106 plusmn 15 115 plusmn 07 155 plusmn 10

Particle diameter in microm mean plusmn U

(k = 2 n = 5) 1114 plusmn 198 1188 plusmn 272 1213 plusmn 60

Soybean DNA yield in ngmg1 mean plusmn U

(k = 2) 07 plusmn 01 (n=5) 579 plusmn 178 (n=3) 729 plusmn 18 (n=2)

Genetic elements detected with screening

pre-spotted plates (Cq value)2

Hmg (Cq 246) Lec (247) CruA (Cq 357) UGP (281) PLD (365) P35S (Cq 333) tNOS (Cq 368) CTP2-EPSPS (Cq 357) Cry1AbAc (Cq 360) PAT (Cq 376)

Lec (Cq 215) Lec (Cq 212) p35S (Cq 212) tNOS (Cq 225) PAT (Cq 348)

GM soybean events detected with event-

specific pre-spotted plates (Cq value)2

40-3-2 (Cq 340) MON87701 (Cq 374) MON89788 (Cq 363) Lec (Cq 246)

NA 40-3-2 (Cq 214) A5547 (Cq 412)

Mass used to prepare T1 (g) 11032 96 108

Nominal target GM mass fraction in T1

(mm ) NA NA 08

1 Results reported here for a sample intake of 200 mg with the in-house validated CTAB method + Genomic-tip 20G purification for soybean (JRC-GEEL) The soybean DNA yield value for the chicken feed was determined by qPCR whereas the yield from the non-GM and GM soybean materials was measured by fluorometry 2 A screening and GM soybean event-specific pre-spotted plate (PSP) was used for these tests NA not applicable k coverage factor U expanded uncertainty

The presence of different species and GM events in the base materials and in a pilot

mixture was tested by using the screening(7) and GM soybean event-specific pre-spotted

plates(8)

The presence of maize and soybean in the chicken feed powder was confirmed by the early

quantification cycle obtained for the high mobility gene (hmg) and lectin (lec) assays Late

amplifications for the UDP-glucose pyrophosphorylase (ugp) phospholipase D (PLD) and

cruciferin A (cruA) genes confirmed the respective presence of potato sunflower and

rapeseed in the chicken feed

The chicken feed powder (labelled bio) also contained traces of genetic markers such as

p35S tNOS EPSPS and Cry1AbAc which indicate a contamination of the chicken feed by

genetically modified plant materials The GM soybean event-specific assays confirmed the

presence of traces of 3 GM soybeans namely 40-3-2 MON87701 and MON89788 The

level of contamination was estimated to be below 001

The final test item was prepared gravimetrically in accordance with ISO 170342016(9) as

follows

The mass of the GM ingredient to add (40-3-2 soybean) was calculated taking into

account the amount of DNA that could be extracted and amplified from the different

materials (Table 1)

The compound sample T1 was mixed in a DynaMIX CM200 for 15 h to improve

homogeneity

After finalisation of the mixing step the powders were filled manually in 20 mL

brown glass vials using lyophilisation inserts manually placed in the bottle necks

Before final closure of the vials air was evacuated in a freeze-dryer and replaced

by argon The vials were finally closed inside the freeze-dryer with the help of a

hydraulic device and then sealed with blue aluminium caps to maintain the inert

atmosphere and to prevent accidental opening during storage and transport

A total of 200 vials containing each at least 5 g of flour were then labelled with a

sample number and the description Sample T1 (chicken feed)

EURL GMFF CT 0217 report

1046

Following the inventory and the selection of vials for future analysis according to a

random stratified sampling scheme the bottles were brought to a storage room for

long-term storage in the dark at 4 plusmn 3 degC

Homogeneity and stability testing of T1 was performed in-house as described in Annex 1

using an event-specific quantification method previously validated by the EURL GMFF

Material T1 was found to be homogeneous for the GM event added (p-value gt 005)

based on a 200 mg sample intake

From the isochronous stability study it was concluded that the test item would be

sufficiently stable under ambient shipment conditions (5 significance level) Stability

was also confirmed during the whole period of the PT between the dispatch of the test

items until the deadline of reporting the results (Annex 1)

JRC-Ispra tested the T1 material and this confirmed the results obtained by JRC-Geel The

average (n = 91) mass fraction of event 40-3-2 measured in T1 was 069 plusmn 004 mm

(U k = 2) which approximated the expected nominal value but may have been

influenced by the characteristics of the different base materials

22 Test item 2

The T2 test item was a new batch of a certified reference material that was not yet

released on the market (Table 2) The bottles of T2 were re-labelled with a unique sample

number as Sample T2 (soybean flour)

Homogeneity and short-term stability of T2 had been previously demonstrated as part of

the certification of the CRM stability monitoring confirmed the stability of T2 during the

running time of the PT (Annex 1)

Table 2 Characteristics of test item 2 (T2)

Characteristic Soybean feed

Type of base material CRM

Origin ERM-BF410dp(6) containing 100 plusmn 06 gkg MON-Oslash4Oslash32-6 soybean produced in 2017 by JRC-Geel

The certificate of ERM-BF410dp warns that a difference (at 95 confidence level)

between the total DNA content in the two powders used for the production of ERM-

BF410dp was found to be significant (due to the different size of non-GM and GM seeds)

and is likely to have an impact when using this CRM Depending on the composition of the

unknown sample real-time PCR measurement results of ERM-BF410dp may differ up to

239 plusmn 11 (average plusmn U) compared to the results of the unknown sample This

difference may depend also on the DNA extraction method selected and both effects may

be additive This observation was confirmed by the participants in this PT who used the

previous 40-3-2 soybean CRM batch (ERM-BF410n) for the calibration of the T2

measurements

EURL GMFF CT 0217 report

1146

3 Instructions to the participants

Participants in this PT were instructed to analyse the two test items (T1 and T2) as

follows

Test Item 1 Chicken feed

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Test Item 2 Soybean flour

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Quantitative results had to be reported in mm as outlined below

Mass GM event [g]

mm = x 100 (1)

Total mass species [g]

Participants were requested

- to use procedures for detectionquantification of the GM events that resemble as

closely as possible the ones used in routine sample analysis

- to take care in ensuring that the DNA extraction method employed is adapted to

the matrix and that the quality of the DNA obtained is suitable for PCR

- to report the quantitative results with two decimal places (eg 064 or 129)

- to follow the general rule that results obtained using a calibrant certified for GM

mass fraction (ie a matrix CRM certified in [x] gkg) can directly be expressed in

mm while results obtained using a calibrant certified for DNA copy number

ratio (eg a plasmid containing both the GM and reference gene target or some

matrix CRMs) need to be converted into mm using a conversion factor(1011)

- to pay attention to the correct estimation and reporting of the measurement

uncertainty and coverage factor used as the uncertainty reported would be

considered in the evaluation of the results using zeta scores

- in case of an unsatisfactory performance to fill in a form indicating the root-cause

analysis and providing evidence demonstrating the effectiveness of the corrective

actions implemented (for NRLs only)

The participants were also informed that the identification information on the participants

in this PT would be kept confidential except for the NRLs that have been appointed in line

with Regulation (EC) No 8822004 their lab codes will be disclosed to DG SANTE for the

purpose of an assessment of their performance

EURL GMFF CT 0217 report

1246

4 Results

41 Participation to the PT

On 31 May 2017 199 laboratories were informed about the upcoming PT EURL-GMFF-CT-

0217 Finally 89 laboratories registered for it and received a random unique lab code

(L01 to L89) Eighty-six laboratories from 39 countries returned results within the

reporting deadline Three non-NRLs did not submit any results two of which (L21 L28)

had not received the samples from customs while L67 did not provide any justification for

not participating

Table 3 shows an overview of the participation to this PT

Table 3 Communication about and participation to the PT 0217

Characteristic of the PT Result

Date of PT announcement 31 May 2017

Deadline for registration 14 June 2017

Date of shipment of samples 3 July 2017

Deadline for result submission 25 August 2017

Number of laboratories informed 199

Number of registered laboratories 89

Registered laboratories that failed to submit their data 3

Number of participating laboratories 86

The participating laboratories fell into the following assigned categories (Table 4)

Thirty-three NRLs designated in line with Regulation (EC) No 8822004 (NRL882)

representing 25 EU Member States (many of them are also NRL120) In addition

Ireland delegated its NRL882 tasks to one of the PT participants Estonia and

Malta were not represented in this PT

Twenty-one NRLs nominated under Regulation (EU) No 1202014 (NRL120) that

are not at the same time NRL under Regulation (EC) No 8822004

Thirty-two laboratories that are not NRL but are appointed by their National

Authority to perform official controls This category includes 10 EU official control

laboratories (OCLs) and 22 laboratories from non-EU countries including Serbia

and Switzerland

Among the countries Germany was represented with 17 laboratories Italy with 6

laboratories and Belgium and Poland with 4 laboratories each all other countries had

between one and three participating laboratories

EURL GMFF CT 0217 report

1346

Table 4 Overview of participants to CT 0217 by country and category

Country Participants NRL8821 NRL120 Non-NRL

AUSTRIA 2 2

BELGIUM 4 3

1

BRAZIL 2

2

BULGARIA 2 1

1

CHILE 1

1

COLOMBIA 1

1

CROATIA 2 1

1

CYPRUS 1 1

CZECH REPUBLIC 1 1

DENMARK 1 1

FINLAND 2 1 1

FRANCE 2 2

GERMANY 17 1 14 2

GREECE 1 1

HUNGARY 2 1

1

INDIA 1

1

ITALY 6 1 2 3

LATVIA 1 1

LEBANON 1

1

LITHUANIA 1 1

LUXEMBOURG 1 1

MEXICO 1

1

NETHERLANDS 2 1 1

PHILIPPINES 1

1

POLAND 4 3 1

PORTUGAL 1 1

ROMANIA 2 1

1

SERBIA 3

3

SINGAPORE 1

1

SLOVAKIA 2 2

SLOVENIA 1 1

SPAIN 2 2

SWEDEN 1 1

SWITZERLAND 2

2

TURKEY 1

1

UKRAINE 3

3

UNITED KINGDOM 3 1 2

UNITED STATES 1

1

VIETNAM 3

3

Total 86 33 21 32 1 No NRL882 from Estonia or Malta participated to this PT

EURL GMFF CT 0217 report

1446

42 Information on the testing provided in the questionnaire

Participants were asked to fill in an online questionnaire (through EUSurvey) on their

testing methodology used for T1 and T2 consisting of a number of mostly multiple-choice

questions A total of 82 laboratories completed the questionnaire including all 33

NRL882 20 out of 21 NRL120 and 29 out of 32 non-NRLs (questionnaires were missing

from L03 L10 L38 and L49)

Table 5 summarises the main answers received whereas Annex 2 shows all answers The

results on GM event identification are reported in Section 43

Table 5 Summary of the main answers provided in the questionnaire of CT 0217

Question (and Question

number) Test Item 1 ndash 40-3-2 Test Item 2 ndash 40-3-2

Test item analysed Yes (801) No (2) Yes (82) No (0)

Reason for lack of analysis (Q1)

Matrix out of scope (1) other practical constraints (1) -

DNA extraction method (Q2) CTAB (41) NucleoSpin Food (10) CTAB (40) NucleoSpin Food (11)

Additional DNA purification method (Q3)

None (51) Ethanol (9) None (52) Ethanol (9)

Number of replicates (Q4) 2 (52) 4 (10) 2 (56) 4 (10)

Approach to test for PCR inhibition (Q5)

OD ratios (40) delta Cq or GM between two dilutions (31)

OD ratios (36) delta Cq or GM between two dilutions (32)

Reason for not testing all events (Q9)

Not applicable (45) below the LOQ (18) Not applicable (61) reagents not available (10)

Approach used (Q6a) Standard curves (68) delta Cq (9) Standard curves (71) delta Cq (9)

Calibrant used (Q6b) CRM JRC-Geel in gkg (71) other RM in copies (4) CRM JRC-Geel in gkg (74) other RM in copies (4)

Taxon-specific endogenous gene (Q6c)

lec-74 bp (65) lec-81 or 118 bp (each 5) lec-74 bp (63) lec-118 bp (6)

Unit of measurement and data expression (Q6d)

Mass (59) copies=mass CRM (13) Mass (61) copies=mass CRM (14)

Amount of DNA (Q6e) 200 ng (30) 100 ng (18) 200 ng (30) 100 ng (21)

LOQ (Q6f) 01 (37) lt01 (34) 01 (39) lt01 (39)

LOQ determination (Q6g) In-house validation (42) current analysis (20) In-house validation (43) current analysis (21)

Uncertainty determination (Q6h)

Precision of replicates (32) in-house validation (27) Precision of replicates (32) in-house validation (30)

1 The numbers shown refer to the number of laboratories that reported the answer The answers that were reported with the two largest frequencies are mentioned

One NRL882 (L73) reported that T1 was out of the scope of the laboratory and one non-

NRL (L43) reported that the T1 matrix was not analysed because of practical constraints

The evaluation of the answers shows that the most commonly employed DNA extraction

method for both T1 and T2 was one based on CTAB with the NucleoSpin Food kit ranking

second No additional purification methods were generally applied The majority of

laboratories analysed two replicate DNA extracts Most laboratories checked the quality of

the DNA extracts by verifying the OD ratios andor running two dilutions a minority of

laboratories performed a PCR inhibition run on 3 or 4 DNA dilutions with a reference gene

For the quantitative analysis the most common approach used was based on two standard

curves however 9 laboratories applied the delta Cq approach One laboratory (L50)

mentioned the use of digital PCR for 40-3-2 soybean quantification in T1 and T2 The

available CRMs from the JRC were used by most laboratories but 5 laboratories used a

non-certified reference material (RM) where values were expressed in GM copy number

ratio (4) or GM mass fraction (1) Lec was used as taxon-specific reference gene by all

laboratories for soybean (mostly the 74 bp version) The majority of laboratories

performed their measurements in the same unit as the certified value of the calibrant used

(gkg) and no conversion factor was applied The LOQ reported was either taken from in-

house validation of the method or determined from the analysis results for this CT In

most cases a LOQ of 01 mm or lower was reported The measurement uncertainty

was either derived from the standard deviation of the measurement replicates or from the

intermediate precision determined in the frame of the single-laboratory validation study

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

646

Executive summary

The European Union Reference Laboratory for Genetically Modified Food and Feed (EURL

GMFF) organised a proficiency test (PT) for National Reference Laboratories (NRLs) to

support the official controls on food and feed in line with Regulation (EC) No 8822004

Other official control laboratories were allowed to participate on a voluntary basis

Two test items were distributed test item 1 (T1) was composed of ground chicken feed

spiked with a mixture of GM soybean event MON-Oslash4Oslash32-6 (40-3-2) and non-GM soybean

and test item 2 (T2) was a soybean flour containing the same GM soybean event 40-3-2

Participants were required to screen T1 and T2 for the presence of three GM soybean

events and to quantify the event that was present with the highest GM mass fraction The

results had to be reported in GM mass fraction (massmass )

Eighty-six participants from 39 countries participated to this PT including 54 NRLs of

which 33 are designated in line with Regulation (EC) No 8822004 (NRL882) and 21 are

nominated in Regulation (EU) No 1202014 to support the EURL GMFF on method

validation (NRL120) as well as 10 other EU official control laboratories

The qualitative results ie the correct identification of the GM event were evaluated and

scored as correct or incorrect The assigned value for the 40-3-2 soybean mass fraction in

both test materials was derived as the robust mean of the data provided by the NRLs

Laboratory performance was primarily evaluated by calculating z scores

The results reported indicate that all participants identified the correct GM event in both

test items All but one NRL obtained a satisfactory performance (z) score for the reported

40-3-2 soybean mass fraction in both test items and the performance of one other NRL

was unacceptable for T1 because they reported the 40-3-2 mass fraction as below their

LOQ Six and two NRLs obtained a questionable z score for T1 and T2 respectively

Considering the results provided by the other participants (non-NRL) three and two non-

NRLs obtained an unsatisfactory z score for T1 and T2 respectively Another three

laboratories obtained a questionable z score for T1 and one laboratory for T2

The laboratories ability to provide results close to the assigned value within their claimed

measurement uncertainty was additionally evaluated by ζ scores Twenty four and 16

laboratories had an unsatisfactory ζ scores for T1 and T2 respectively Unsatisfactory ζ

scores were mainly the consequence of an underestimated or not reported measurement

uncertainty Guidance is provided for correctly estimating the measurement uncertainty of

analytical results

A root-cause analysis will be requested from NRLs with an unsatisfactory z score result in

this PT and will be followed-up

EURL GMFF CT 0217 report

746

1 Introduction

The Joint Research Centre (JRC) of the European Commission was established as European

Union Reference Laboratory for GM Food and Feed (EURL GMFF) by Regulations (EC) No

18292003(1) and (EC) No 8822004(2) Regulation (EC) No 8822004 also requires

Member States to designate National Reference Laboratories (NRL882) for each EURL

coordinating activities for the official control of compliance with food and feed law The

analytical methods used for these controls have been validated by the EURL GMFF as

required by Regulation (EC) No 18292003 and for this task the EURL GMFF is supported

by NRLs listed in Regulation (EU) No 1202014(3) (NRL120 several of them are also

NRL882) The Member States of the European Union may also appoint other laboratories

(non-NRLs) to perform the official controls on food and feed

It is crucial that official control laboratories can accurately and reliably determine the GM

content in food and feed samples Regulation (EC) No 18292003 established a threshold

for labelling of food and feed products containing genetically modified material that is

authorised in the EU (09 ) Furthermore Regulation (EU) No 6192011(4) introduced a

minimum performance limit (01 mm ) for detecting the accidental presence in feed of

genetically modified material with pending or expired authorisation status Compliance

with these values is verified by the Member States of the European Union in the official

control of food and feed

The EURL GMFF is tasked with the organisation of proficiency tests (called comparative

tests or CT in the GMO legislation(2)) to foster the correct application of the analytical

methods available for the official controls The EURL GMFF is operating under a quality

management system which is accredited according to ISOIEC 17043(5) for the

organisation of proficiency testing

This report summarises the results obtained in a PT organised by the EURL GMFF in 2017

(CT 0217) Participation in such PTs is mandatory for NRL882 recommended for

NRL120 and open to other official control laboratories

EURL GMFF CT 0217 report

846

2 Test items

The test items used in this PT were prepared and characterised at JRC-Geel

21 Test item 1

The T1 test item was prepared from base materials that were characterised before their

use (Table 1) The base materials employed for the preparation of T1 were chicken feed

(AVEVE for biological agriculture according to EC 8342007 and EC 8892008) the ERM-

BF410ak as non GM soybean powder and the ERM-BF410bp(6) containing the MON-Oslash4Oslash32-

6 event (hereafter named 40-3-2) as spiking material (Table 1)

The chicken feed was composed according to the label of Bio maize kernels Bio soybean

oil-cake Bio-sunflower oil-cake Bio wheat Bio barley maize gluten potato proteins

calcium monophosphate calcium carbonate Bio soybean oil and sodium chloride The

analytical composition indicated a content of 175 protein 5 fat 13 ashes 5

cellulose 033 methionine 078 lysine 37 calcium 052 phosphorus and

01 sodium The chicken feed was milled using a cryo-grinding vibrating mill (Palla mill

KHD Humboldt-Wedag Koumlln DE) and sieved with a 500 microm stainless steel mesh on a

sieving machine equipped with an ultrasonic sieving aid (Russel Finex London UK) The

remaining powder was mixed in a DynaMIX CM200 (WAB Muttenz CH) for 1 h to

homogenise the distribution of the different types of seed tissues

The residual water mass fractions for the chicken feed powder and the powders of the

certified reference materials ERM-BF410ak and ERM-BF410bp were measured by

volumetric Karl Fischer titration (758 KFD Titrino Metrohm Herisau CH) The results

showed that the powders were sufficiently dry to perform the dry mixing and did not

require an additional drying step

The particle size distribution of the powders was measured using laser diffraction (PSA

Sympatec Clausthal-Zellerfeld DE) It was concluded that the particle size distribution of

these powders was sufficiently similar to allow subsequent preparations of mixtures

The amount and the quality of the DNA extracted from the chicken feed powder the non-

GM soybean flour and the GM spiking material were verified by UV spectrometry

fluorometry and gel electrophoresis A CTAB-tip 20G method (Qiagen Hilden Germany)

optimised for soybean was chosen with a sample intake of 200 mg because it yielded a

sufficient amount of DNA of PCR-grade quality from the base materials DNA extracted

with the in-house CTAB method was tested for PCR inhibition between 40 ngmicroL to 02

ngmicroL with a lectin qPCR assay (5 microL per PCR) and did not show any inhibition (Cq values

were very close to the theoretical Cq values) The PCR efficiencies ranged from 96 to 98

with a coefficient of determination (R2) between 099 and 100 confirming the absence

of significant amounts of PCR inhibitors in the extracts

The CTAB method yielded a sufficient amount of DNA of PCR-grade quality from both non-

GM and GM base materials

The level of fragmentation of the extracted DNA was investigated by 10 agarose gel

electrophoresis A smear from plusmn 12 to 1 kbp could be clearly seen in the chicken feed

DNA indicating some level of fragmentation of the extracted DNA while the DNA

extracted from the soybean materials migrated as a high molecular weight band (above 12

kbp) The amount of soybean DNA that could be extracted and amplified from the chicken

feed powder was determined by qPCR with a lectin assay using DNA from a soybean CRM

as calibrant this amount appeared to be rather low (lt1 of total DNA) The yield of

amplifiable DNA per mg of chicken feed powder and the yield of DNA measured by

PicoGreen for the soybean materials composed of pure soybean were taken into account

to calculate the amount of chicken feed powder non-GM soybean and GM soybean to be

mixed to obtain a target value of approximately 08 mm event 40-3-2 in T1

EURL GMFF CT 0217 report

946

Table 1 Characteristics of the base materials used for preparation of test item 1 (T1)

Characteristic Chicken Feed Non-GM

Soybean Flour

40-3-2

Soybean Flour

Type of base material Scratch grains CRM CRM

Origin AVEVE (Belgium) ERM-BF410ak ERM-BF410bp

Grinding method Cryo-grinding vibrating mill Used as such Used as such

Mixing method DynaMIX CM200 (WAB Muttenz CH)

Water content in gkg mean plusmn U

(k = 2 n = 3) 106 plusmn 15 115 plusmn 07 155 plusmn 10

Particle diameter in microm mean plusmn U

(k = 2 n = 5) 1114 plusmn 198 1188 plusmn 272 1213 plusmn 60

Soybean DNA yield in ngmg1 mean plusmn U

(k = 2) 07 plusmn 01 (n=5) 579 plusmn 178 (n=3) 729 plusmn 18 (n=2)

Genetic elements detected with screening

pre-spotted plates (Cq value)2

Hmg (Cq 246) Lec (247) CruA (Cq 357) UGP (281) PLD (365) P35S (Cq 333) tNOS (Cq 368) CTP2-EPSPS (Cq 357) Cry1AbAc (Cq 360) PAT (Cq 376)

Lec (Cq 215) Lec (Cq 212) p35S (Cq 212) tNOS (Cq 225) PAT (Cq 348)

GM soybean events detected with event-

specific pre-spotted plates (Cq value)2

40-3-2 (Cq 340) MON87701 (Cq 374) MON89788 (Cq 363) Lec (Cq 246)

NA 40-3-2 (Cq 214) A5547 (Cq 412)

Mass used to prepare T1 (g) 11032 96 108

Nominal target GM mass fraction in T1

(mm ) NA NA 08

1 Results reported here for a sample intake of 200 mg with the in-house validated CTAB method + Genomic-tip 20G purification for soybean (JRC-GEEL) The soybean DNA yield value for the chicken feed was determined by qPCR whereas the yield from the non-GM and GM soybean materials was measured by fluorometry 2 A screening and GM soybean event-specific pre-spotted plate (PSP) was used for these tests NA not applicable k coverage factor U expanded uncertainty

The presence of different species and GM events in the base materials and in a pilot

mixture was tested by using the screening(7) and GM soybean event-specific pre-spotted

plates(8)

The presence of maize and soybean in the chicken feed powder was confirmed by the early

quantification cycle obtained for the high mobility gene (hmg) and lectin (lec) assays Late

amplifications for the UDP-glucose pyrophosphorylase (ugp) phospholipase D (PLD) and

cruciferin A (cruA) genes confirmed the respective presence of potato sunflower and

rapeseed in the chicken feed

The chicken feed powder (labelled bio) also contained traces of genetic markers such as

p35S tNOS EPSPS and Cry1AbAc which indicate a contamination of the chicken feed by

genetically modified plant materials The GM soybean event-specific assays confirmed the

presence of traces of 3 GM soybeans namely 40-3-2 MON87701 and MON89788 The

level of contamination was estimated to be below 001

The final test item was prepared gravimetrically in accordance with ISO 170342016(9) as

follows

The mass of the GM ingredient to add (40-3-2 soybean) was calculated taking into

account the amount of DNA that could be extracted and amplified from the different

materials (Table 1)

The compound sample T1 was mixed in a DynaMIX CM200 for 15 h to improve

homogeneity

After finalisation of the mixing step the powders were filled manually in 20 mL

brown glass vials using lyophilisation inserts manually placed in the bottle necks

Before final closure of the vials air was evacuated in a freeze-dryer and replaced

by argon The vials were finally closed inside the freeze-dryer with the help of a

hydraulic device and then sealed with blue aluminium caps to maintain the inert

atmosphere and to prevent accidental opening during storage and transport

A total of 200 vials containing each at least 5 g of flour were then labelled with a

sample number and the description Sample T1 (chicken feed)

EURL GMFF CT 0217 report

1046

Following the inventory and the selection of vials for future analysis according to a

random stratified sampling scheme the bottles were brought to a storage room for

long-term storage in the dark at 4 plusmn 3 degC

Homogeneity and stability testing of T1 was performed in-house as described in Annex 1

using an event-specific quantification method previously validated by the EURL GMFF

Material T1 was found to be homogeneous for the GM event added (p-value gt 005)

based on a 200 mg sample intake

From the isochronous stability study it was concluded that the test item would be

sufficiently stable under ambient shipment conditions (5 significance level) Stability

was also confirmed during the whole period of the PT between the dispatch of the test

items until the deadline of reporting the results (Annex 1)

JRC-Ispra tested the T1 material and this confirmed the results obtained by JRC-Geel The

average (n = 91) mass fraction of event 40-3-2 measured in T1 was 069 plusmn 004 mm

(U k = 2) which approximated the expected nominal value but may have been

influenced by the characteristics of the different base materials

22 Test item 2

The T2 test item was a new batch of a certified reference material that was not yet

released on the market (Table 2) The bottles of T2 were re-labelled with a unique sample

number as Sample T2 (soybean flour)

Homogeneity and short-term stability of T2 had been previously demonstrated as part of

the certification of the CRM stability monitoring confirmed the stability of T2 during the

running time of the PT (Annex 1)

Table 2 Characteristics of test item 2 (T2)

Characteristic Soybean feed

Type of base material CRM

Origin ERM-BF410dp(6) containing 100 plusmn 06 gkg MON-Oslash4Oslash32-6 soybean produced in 2017 by JRC-Geel

The certificate of ERM-BF410dp warns that a difference (at 95 confidence level)

between the total DNA content in the two powders used for the production of ERM-

BF410dp was found to be significant (due to the different size of non-GM and GM seeds)

and is likely to have an impact when using this CRM Depending on the composition of the

unknown sample real-time PCR measurement results of ERM-BF410dp may differ up to

239 plusmn 11 (average plusmn U) compared to the results of the unknown sample This

difference may depend also on the DNA extraction method selected and both effects may

be additive This observation was confirmed by the participants in this PT who used the

previous 40-3-2 soybean CRM batch (ERM-BF410n) for the calibration of the T2

measurements

EURL GMFF CT 0217 report

1146

3 Instructions to the participants

Participants in this PT were instructed to analyse the two test items (T1 and T2) as

follows

Test Item 1 Chicken feed

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Test Item 2 Soybean flour

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Quantitative results had to be reported in mm as outlined below

Mass GM event [g]

mm = x 100 (1)

Total mass species [g]

Participants were requested

- to use procedures for detectionquantification of the GM events that resemble as

closely as possible the ones used in routine sample analysis

- to take care in ensuring that the DNA extraction method employed is adapted to

the matrix and that the quality of the DNA obtained is suitable for PCR

- to report the quantitative results with two decimal places (eg 064 or 129)

- to follow the general rule that results obtained using a calibrant certified for GM

mass fraction (ie a matrix CRM certified in [x] gkg) can directly be expressed in

mm while results obtained using a calibrant certified for DNA copy number

ratio (eg a plasmid containing both the GM and reference gene target or some

matrix CRMs) need to be converted into mm using a conversion factor(1011)

- to pay attention to the correct estimation and reporting of the measurement

uncertainty and coverage factor used as the uncertainty reported would be

considered in the evaluation of the results using zeta scores

- in case of an unsatisfactory performance to fill in a form indicating the root-cause

analysis and providing evidence demonstrating the effectiveness of the corrective

actions implemented (for NRLs only)

The participants were also informed that the identification information on the participants

in this PT would be kept confidential except for the NRLs that have been appointed in line

with Regulation (EC) No 8822004 their lab codes will be disclosed to DG SANTE for the

purpose of an assessment of their performance

EURL GMFF CT 0217 report

1246

4 Results

41 Participation to the PT

On 31 May 2017 199 laboratories were informed about the upcoming PT EURL-GMFF-CT-

0217 Finally 89 laboratories registered for it and received a random unique lab code

(L01 to L89) Eighty-six laboratories from 39 countries returned results within the

reporting deadline Three non-NRLs did not submit any results two of which (L21 L28)

had not received the samples from customs while L67 did not provide any justification for

not participating

Table 3 shows an overview of the participation to this PT

Table 3 Communication about and participation to the PT 0217

Characteristic of the PT Result

Date of PT announcement 31 May 2017

Deadline for registration 14 June 2017

Date of shipment of samples 3 July 2017

Deadline for result submission 25 August 2017

Number of laboratories informed 199

Number of registered laboratories 89

Registered laboratories that failed to submit their data 3

Number of participating laboratories 86

The participating laboratories fell into the following assigned categories (Table 4)

Thirty-three NRLs designated in line with Regulation (EC) No 8822004 (NRL882)

representing 25 EU Member States (many of them are also NRL120) In addition

Ireland delegated its NRL882 tasks to one of the PT participants Estonia and

Malta were not represented in this PT

Twenty-one NRLs nominated under Regulation (EU) No 1202014 (NRL120) that

are not at the same time NRL under Regulation (EC) No 8822004

Thirty-two laboratories that are not NRL but are appointed by their National

Authority to perform official controls This category includes 10 EU official control

laboratories (OCLs) and 22 laboratories from non-EU countries including Serbia

and Switzerland

Among the countries Germany was represented with 17 laboratories Italy with 6

laboratories and Belgium and Poland with 4 laboratories each all other countries had

between one and three participating laboratories

EURL GMFF CT 0217 report

1346

Table 4 Overview of participants to CT 0217 by country and category

Country Participants NRL8821 NRL120 Non-NRL

AUSTRIA 2 2

BELGIUM 4 3

1

BRAZIL 2

2

BULGARIA 2 1

1

CHILE 1

1

COLOMBIA 1

1

CROATIA 2 1

1

CYPRUS 1 1

CZECH REPUBLIC 1 1

DENMARK 1 1

FINLAND 2 1 1

FRANCE 2 2

GERMANY 17 1 14 2

GREECE 1 1

HUNGARY 2 1

1

INDIA 1

1

ITALY 6 1 2 3

LATVIA 1 1

LEBANON 1

1

LITHUANIA 1 1

LUXEMBOURG 1 1

MEXICO 1

1

NETHERLANDS 2 1 1

PHILIPPINES 1

1

POLAND 4 3 1

PORTUGAL 1 1

ROMANIA 2 1

1

SERBIA 3

3

SINGAPORE 1

1

SLOVAKIA 2 2

SLOVENIA 1 1

SPAIN 2 2

SWEDEN 1 1

SWITZERLAND 2

2

TURKEY 1

1

UKRAINE 3

3

UNITED KINGDOM 3 1 2

UNITED STATES 1

1

VIETNAM 3

3

Total 86 33 21 32 1 No NRL882 from Estonia or Malta participated to this PT

EURL GMFF CT 0217 report

1446

42 Information on the testing provided in the questionnaire

Participants were asked to fill in an online questionnaire (through EUSurvey) on their

testing methodology used for T1 and T2 consisting of a number of mostly multiple-choice

questions A total of 82 laboratories completed the questionnaire including all 33

NRL882 20 out of 21 NRL120 and 29 out of 32 non-NRLs (questionnaires were missing

from L03 L10 L38 and L49)

Table 5 summarises the main answers received whereas Annex 2 shows all answers The

results on GM event identification are reported in Section 43

Table 5 Summary of the main answers provided in the questionnaire of CT 0217

Question (and Question

number) Test Item 1 ndash 40-3-2 Test Item 2 ndash 40-3-2

Test item analysed Yes (801) No (2) Yes (82) No (0)

Reason for lack of analysis (Q1)

Matrix out of scope (1) other practical constraints (1) -

DNA extraction method (Q2) CTAB (41) NucleoSpin Food (10) CTAB (40) NucleoSpin Food (11)

Additional DNA purification method (Q3)

None (51) Ethanol (9) None (52) Ethanol (9)

Number of replicates (Q4) 2 (52) 4 (10) 2 (56) 4 (10)

Approach to test for PCR inhibition (Q5)

OD ratios (40) delta Cq or GM between two dilutions (31)

OD ratios (36) delta Cq or GM between two dilutions (32)

Reason for not testing all events (Q9)

Not applicable (45) below the LOQ (18) Not applicable (61) reagents not available (10)

Approach used (Q6a) Standard curves (68) delta Cq (9) Standard curves (71) delta Cq (9)

Calibrant used (Q6b) CRM JRC-Geel in gkg (71) other RM in copies (4) CRM JRC-Geel in gkg (74) other RM in copies (4)

Taxon-specific endogenous gene (Q6c)

lec-74 bp (65) lec-81 or 118 bp (each 5) lec-74 bp (63) lec-118 bp (6)

Unit of measurement and data expression (Q6d)

Mass (59) copies=mass CRM (13) Mass (61) copies=mass CRM (14)

Amount of DNA (Q6e) 200 ng (30) 100 ng (18) 200 ng (30) 100 ng (21)

LOQ (Q6f) 01 (37) lt01 (34) 01 (39) lt01 (39)

LOQ determination (Q6g) In-house validation (42) current analysis (20) In-house validation (43) current analysis (21)

Uncertainty determination (Q6h)

Precision of replicates (32) in-house validation (27) Precision of replicates (32) in-house validation (30)

1 The numbers shown refer to the number of laboratories that reported the answer The answers that were reported with the two largest frequencies are mentioned

One NRL882 (L73) reported that T1 was out of the scope of the laboratory and one non-

NRL (L43) reported that the T1 matrix was not analysed because of practical constraints

The evaluation of the answers shows that the most commonly employed DNA extraction

method for both T1 and T2 was one based on CTAB with the NucleoSpin Food kit ranking

second No additional purification methods were generally applied The majority of

laboratories analysed two replicate DNA extracts Most laboratories checked the quality of

the DNA extracts by verifying the OD ratios andor running two dilutions a minority of

laboratories performed a PCR inhibition run on 3 or 4 DNA dilutions with a reference gene

For the quantitative analysis the most common approach used was based on two standard

curves however 9 laboratories applied the delta Cq approach One laboratory (L50)

mentioned the use of digital PCR for 40-3-2 soybean quantification in T1 and T2 The

available CRMs from the JRC were used by most laboratories but 5 laboratories used a

non-certified reference material (RM) where values were expressed in GM copy number

ratio (4) or GM mass fraction (1) Lec was used as taxon-specific reference gene by all

laboratories for soybean (mostly the 74 bp version) The majority of laboratories

performed their measurements in the same unit as the certified value of the calibrant used

(gkg) and no conversion factor was applied The LOQ reported was either taken from in-

house validation of the method or determined from the analysis results for this CT In

most cases a LOQ of 01 mm or lower was reported The measurement uncertainty

was either derived from the standard deviation of the measurement replicates or from the

intermediate precision determined in the frame of the single-laboratory validation study

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

746

1 Introduction

The Joint Research Centre (JRC) of the European Commission was established as European

Union Reference Laboratory for GM Food and Feed (EURL GMFF) by Regulations (EC) No

18292003(1) and (EC) No 8822004(2) Regulation (EC) No 8822004 also requires

Member States to designate National Reference Laboratories (NRL882) for each EURL

coordinating activities for the official control of compliance with food and feed law The

analytical methods used for these controls have been validated by the EURL GMFF as

required by Regulation (EC) No 18292003 and for this task the EURL GMFF is supported

by NRLs listed in Regulation (EU) No 1202014(3) (NRL120 several of them are also

NRL882) The Member States of the European Union may also appoint other laboratories

(non-NRLs) to perform the official controls on food and feed

It is crucial that official control laboratories can accurately and reliably determine the GM

content in food and feed samples Regulation (EC) No 18292003 established a threshold

for labelling of food and feed products containing genetically modified material that is

authorised in the EU (09 ) Furthermore Regulation (EU) No 6192011(4) introduced a

minimum performance limit (01 mm ) for detecting the accidental presence in feed of

genetically modified material with pending or expired authorisation status Compliance

with these values is verified by the Member States of the European Union in the official

control of food and feed

The EURL GMFF is tasked with the organisation of proficiency tests (called comparative

tests or CT in the GMO legislation(2)) to foster the correct application of the analytical

methods available for the official controls The EURL GMFF is operating under a quality

management system which is accredited according to ISOIEC 17043(5) for the

organisation of proficiency testing

This report summarises the results obtained in a PT organised by the EURL GMFF in 2017

(CT 0217) Participation in such PTs is mandatory for NRL882 recommended for

NRL120 and open to other official control laboratories

EURL GMFF CT 0217 report

846

2 Test items

The test items used in this PT were prepared and characterised at JRC-Geel

21 Test item 1

The T1 test item was prepared from base materials that were characterised before their

use (Table 1) The base materials employed for the preparation of T1 were chicken feed

(AVEVE for biological agriculture according to EC 8342007 and EC 8892008) the ERM-

BF410ak as non GM soybean powder and the ERM-BF410bp(6) containing the MON-Oslash4Oslash32-

6 event (hereafter named 40-3-2) as spiking material (Table 1)

The chicken feed was composed according to the label of Bio maize kernels Bio soybean

oil-cake Bio-sunflower oil-cake Bio wheat Bio barley maize gluten potato proteins

calcium monophosphate calcium carbonate Bio soybean oil and sodium chloride The

analytical composition indicated a content of 175 protein 5 fat 13 ashes 5

cellulose 033 methionine 078 lysine 37 calcium 052 phosphorus and

01 sodium The chicken feed was milled using a cryo-grinding vibrating mill (Palla mill

KHD Humboldt-Wedag Koumlln DE) and sieved with a 500 microm stainless steel mesh on a

sieving machine equipped with an ultrasonic sieving aid (Russel Finex London UK) The

remaining powder was mixed in a DynaMIX CM200 (WAB Muttenz CH) for 1 h to

homogenise the distribution of the different types of seed tissues

The residual water mass fractions for the chicken feed powder and the powders of the

certified reference materials ERM-BF410ak and ERM-BF410bp were measured by

volumetric Karl Fischer titration (758 KFD Titrino Metrohm Herisau CH) The results

showed that the powders were sufficiently dry to perform the dry mixing and did not

require an additional drying step

The particle size distribution of the powders was measured using laser diffraction (PSA

Sympatec Clausthal-Zellerfeld DE) It was concluded that the particle size distribution of

these powders was sufficiently similar to allow subsequent preparations of mixtures

The amount and the quality of the DNA extracted from the chicken feed powder the non-

GM soybean flour and the GM spiking material were verified by UV spectrometry

fluorometry and gel electrophoresis A CTAB-tip 20G method (Qiagen Hilden Germany)

optimised for soybean was chosen with a sample intake of 200 mg because it yielded a

sufficient amount of DNA of PCR-grade quality from the base materials DNA extracted

with the in-house CTAB method was tested for PCR inhibition between 40 ngmicroL to 02

ngmicroL with a lectin qPCR assay (5 microL per PCR) and did not show any inhibition (Cq values

were very close to the theoretical Cq values) The PCR efficiencies ranged from 96 to 98

with a coefficient of determination (R2) between 099 and 100 confirming the absence

of significant amounts of PCR inhibitors in the extracts

The CTAB method yielded a sufficient amount of DNA of PCR-grade quality from both non-

GM and GM base materials

The level of fragmentation of the extracted DNA was investigated by 10 agarose gel

electrophoresis A smear from plusmn 12 to 1 kbp could be clearly seen in the chicken feed

DNA indicating some level of fragmentation of the extracted DNA while the DNA

extracted from the soybean materials migrated as a high molecular weight band (above 12

kbp) The amount of soybean DNA that could be extracted and amplified from the chicken

feed powder was determined by qPCR with a lectin assay using DNA from a soybean CRM

as calibrant this amount appeared to be rather low (lt1 of total DNA) The yield of

amplifiable DNA per mg of chicken feed powder and the yield of DNA measured by

PicoGreen for the soybean materials composed of pure soybean were taken into account

to calculate the amount of chicken feed powder non-GM soybean and GM soybean to be

mixed to obtain a target value of approximately 08 mm event 40-3-2 in T1

EURL GMFF CT 0217 report

946

Table 1 Characteristics of the base materials used for preparation of test item 1 (T1)

Characteristic Chicken Feed Non-GM

Soybean Flour

40-3-2

Soybean Flour

Type of base material Scratch grains CRM CRM

Origin AVEVE (Belgium) ERM-BF410ak ERM-BF410bp

Grinding method Cryo-grinding vibrating mill Used as such Used as such

Mixing method DynaMIX CM200 (WAB Muttenz CH)

Water content in gkg mean plusmn U

(k = 2 n = 3) 106 plusmn 15 115 plusmn 07 155 plusmn 10

Particle diameter in microm mean plusmn U

(k = 2 n = 5) 1114 plusmn 198 1188 plusmn 272 1213 plusmn 60

Soybean DNA yield in ngmg1 mean plusmn U

(k = 2) 07 plusmn 01 (n=5) 579 plusmn 178 (n=3) 729 plusmn 18 (n=2)

Genetic elements detected with screening

pre-spotted plates (Cq value)2

Hmg (Cq 246) Lec (247) CruA (Cq 357) UGP (281) PLD (365) P35S (Cq 333) tNOS (Cq 368) CTP2-EPSPS (Cq 357) Cry1AbAc (Cq 360) PAT (Cq 376)

Lec (Cq 215) Lec (Cq 212) p35S (Cq 212) tNOS (Cq 225) PAT (Cq 348)

GM soybean events detected with event-

specific pre-spotted plates (Cq value)2

40-3-2 (Cq 340) MON87701 (Cq 374) MON89788 (Cq 363) Lec (Cq 246)

NA 40-3-2 (Cq 214) A5547 (Cq 412)

Mass used to prepare T1 (g) 11032 96 108

Nominal target GM mass fraction in T1

(mm ) NA NA 08

1 Results reported here for a sample intake of 200 mg with the in-house validated CTAB method + Genomic-tip 20G purification for soybean (JRC-GEEL) The soybean DNA yield value for the chicken feed was determined by qPCR whereas the yield from the non-GM and GM soybean materials was measured by fluorometry 2 A screening and GM soybean event-specific pre-spotted plate (PSP) was used for these tests NA not applicable k coverage factor U expanded uncertainty

The presence of different species and GM events in the base materials and in a pilot

mixture was tested by using the screening(7) and GM soybean event-specific pre-spotted

plates(8)

The presence of maize and soybean in the chicken feed powder was confirmed by the early

quantification cycle obtained for the high mobility gene (hmg) and lectin (lec) assays Late

amplifications for the UDP-glucose pyrophosphorylase (ugp) phospholipase D (PLD) and

cruciferin A (cruA) genes confirmed the respective presence of potato sunflower and

rapeseed in the chicken feed

The chicken feed powder (labelled bio) also contained traces of genetic markers such as

p35S tNOS EPSPS and Cry1AbAc which indicate a contamination of the chicken feed by

genetically modified plant materials The GM soybean event-specific assays confirmed the

presence of traces of 3 GM soybeans namely 40-3-2 MON87701 and MON89788 The

level of contamination was estimated to be below 001

The final test item was prepared gravimetrically in accordance with ISO 170342016(9) as

follows

The mass of the GM ingredient to add (40-3-2 soybean) was calculated taking into

account the amount of DNA that could be extracted and amplified from the different

materials (Table 1)

The compound sample T1 was mixed in a DynaMIX CM200 for 15 h to improve

homogeneity

After finalisation of the mixing step the powders were filled manually in 20 mL

brown glass vials using lyophilisation inserts manually placed in the bottle necks

Before final closure of the vials air was evacuated in a freeze-dryer and replaced

by argon The vials were finally closed inside the freeze-dryer with the help of a

hydraulic device and then sealed with blue aluminium caps to maintain the inert

atmosphere and to prevent accidental opening during storage and transport

A total of 200 vials containing each at least 5 g of flour were then labelled with a

sample number and the description Sample T1 (chicken feed)

EURL GMFF CT 0217 report

1046

Following the inventory and the selection of vials for future analysis according to a

random stratified sampling scheme the bottles were brought to a storage room for

long-term storage in the dark at 4 plusmn 3 degC

Homogeneity and stability testing of T1 was performed in-house as described in Annex 1

using an event-specific quantification method previously validated by the EURL GMFF

Material T1 was found to be homogeneous for the GM event added (p-value gt 005)

based on a 200 mg sample intake

From the isochronous stability study it was concluded that the test item would be

sufficiently stable under ambient shipment conditions (5 significance level) Stability

was also confirmed during the whole period of the PT between the dispatch of the test

items until the deadline of reporting the results (Annex 1)

JRC-Ispra tested the T1 material and this confirmed the results obtained by JRC-Geel The

average (n = 91) mass fraction of event 40-3-2 measured in T1 was 069 plusmn 004 mm

(U k = 2) which approximated the expected nominal value but may have been

influenced by the characteristics of the different base materials

22 Test item 2

The T2 test item was a new batch of a certified reference material that was not yet

released on the market (Table 2) The bottles of T2 were re-labelled with a unique sample

number as Sample T2 (soybean flour)

Homogeneity and short-term stability of T2 had been previously demonstrated as part of

the certification of the CRM stability monitoring confirmed the stability of T2 during the

running time of the PT (Annex 1)

Table 2 Characteristics of test item 2 (T2)

Characteristic Soybean feed

Type of base material CRM

Origin ERM-BF410dp(6) containing 100 plusmn 06 gkg MON-Oslash4Oslash32-6 soybean produced in 2017 by JRC-Geel

The certificate of ERM-BF410dp warns that a difference (at 95 confidence level)

between the total DNA content in the two powders used for the production of ERM-

BF410dp was found to be significant (due to the different size of non-GM and GM seeds)

and is likely to have an impact when using this CRM Depending on the composition of the

unknown sample real-time PCR measurement results of ERM-BF410dp may differ up to

239 plusmn 11 (average plusmn U) compared to the results of the unknown sample This

difference may depend also on the DNA extraction method selected and both effects may

be additive This observation was confirmed by the participants in this PT who used the

previous 40-3-2 soybean CRM batch (ERM-BF410n) for the calibration of the T2

measurements

EURL GMFF CT 0217 report

1146

3 Instructions to the participants

Participants in this PT were instructed to analyse the two test items (T1 and T2) as

follows

Test Item 1 Chicken feed

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Test Item 2 Soybean flour

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Quantitative results had to be reported in mm as outlined below

Mass GM event [g]

mm = x 100 (1)

Total mass species [g]

Participants were requested

- to use procedures for detectionquantification of the GM events that resemble as

closely as possible the ones used in routine sample analysis

- to take care in ensuring that the DNA extraction method employed is adapted to

the matrix and that the quality of the DNA obtained is suitable for PCR

- to report the quantitative results with two decimal places (eg 064 or 129)

- to follow the general rule that results obtained using a calibrant certified for GM

mass fraction (ie a matrix CRM certified in [x] gkg) can directly be expressed in

mm while results obtained using a calibrant certified for DNA copy number

ratio (eg a plasmid containing both the GM and reference gene target or some

matrix CRMs) need to be converted into mm using a conversion factor(1011)

- to pay attention to the correct estimation and reporting of the measurement

uncertainty and coverage factor used as the uncertainty reported would be

considered in the evaluation of the results using zeta scores

- in case of an unsatisfactory performance to fill in a form indicating the root-cause

analysis and providing evidence demonstrating the effectiveness of the corrective

actions implemented (for NRLs only)

The participants were also informed that the identification information on the participants

in this PT would be kept confidential except for the NRLs that have been appointed in line

with Regulation (EC) No 8822004 their lab codes will be disclosed to DG SANTE for the

purpose of an assessment of their performance

EURL GMFF CT 0217 report

1246

4 Results

41 Participation to the PT

On 31 May 2017 199 laboratories were informed about the upcoming PT EURL-GMFF-CT-

0217 Finally 89 laboratories registered for it and received a random unique lab code

(L01 to L89) Eighty-six laboratories from 39 countries returned results within the

reporting deadline Three non-NRLs did not submit any results two of which (L21 L28)

had not received the samples from customs while L67 did not provide any justification for

not participating

Table 3 shows an overview of the participation to this PT

Table 3 Communication about and participation to the PT 0217

Characteristic of the PT Result

Date of PT announcement 31 May 2017

Deadline for registration 14 June 2017

Date of shipment of samples 3 July 2017

Deadline for result submission 25 August 2017

Number of laboratories informed 199

Number of registered laboratories 89

Registered laboratories that failed to submit their data 3

Number of participating laboratories 86

The participating laboratories fell into the following assigned categories (Table 4)

Thirty-three NRLs designated in line with Regulation (EC) No 8822004 (NRL882)

representing 25 EU Member States (many of them are also NRL120) In addition

Ireland delegated its NRL882 tasks to one of the PT participants Estonia and

Malta were not represented in this PT

Twenty-one NRLs nominated under Regulation (EU) No 1202014 (NRL120) that

are not at the same time NRL under Regulation (EC) No 8822004

Thirty-two laboratories that are not NRL but are appointed by their National

Authority to perform official controls This category includes 10 EU official control

laboratories (OCLs) and 22 laboratories from non-EU countries including Serbia

and Switzerland

Among the countries Germany was represented with 17 laboratories Italy with 6

laboratories and Belgium and Poland with 4 laboratories each all other countries had

between one and three participating laboratories

EURL GMFF CT 0217 report

1346

Table 4 Overview of participants to CT 0217 by country and category

Country Participants NRL8821 NRL120 Non-NRL

AUSTRIA 2 2

BELGIUM 4 3

1

BRAZIL 2

2

BULGARIA 2 1

1

CHILE 1

1

COLOMBIA 1

1

CROATIA 2 1

1

CYPRUS 1 1

CZECH REPUBLIC 1 1

DENMARK 1 1

FINLAND 2 1 1

FRANCE 2 2

GERMANY 17 1 14 2

GREECE 1 1

HUNGARY 2 1

1

INDIA 1

1

ITALY 6 1 2 3

LATVIA 1 1

LEBANON 1

1

LITHUANIA 1 1

LUXEMBOURG 1 1

MEXICO 1

1

NETHERLANDS 2 1 1

PHILIPPINES 1

1

POLAND 4 3 1

PORTUGAL 1 1

ROMANIA 2 1

1

SERBIA 3

3

SINGAPORE 1

1

SLOVAKIA 2 2

SLOVENIA 1 1

SPAIN 2 2

SWEDEN 1 1

SWITZERLAND 2

2

TURKEY 1

1

UKRAINE 3

3

UNITED KINGDOM 3 1 2

UNITED STATES 1

1

VIETNAM 3

3

Total 86 33 21 32 1 No NRL882 from Estonia or Malta participated to this PT

EURL GMFF CT 0217 report

1446

42 Information on the testing provided in the questionnaire

Participants were asked to fill in an online questionnaire (through EUSurvey) on their

testing methodology used for T1 and T2 consisting of a number of mostly multiple-choice

questions A total of 82 laboratories completed the questionnaire including all 33

NRL882 20 out of 21 NRL120 and 29 out of 32 non-NRLs (questionnaires were missing

from L03 L10 L38 and L49)

Table 5 summarises the main answers received whereas Annex 2 shows all answers The

results on GM event identification are reported in Section 43

Table 5 Summary of the main answers provided in the questionnaire of CT 0217

Question (and Question

number) Test Item 1 ndash 40-3-2 Test Item 2 ndash 40-3-2

Test item analysed Yes (801) No (2) Yes (82) No (0)

Reason for lack of analysis (Q1)

Matrix out of scope (1) other practical constraints (1) -

DNA extraction method (Q2) CTAB (41) NucleoSpin Food (10) CTAB (40) NucleoSpin Food (11)

Additional DNA purification method (Q3)

None (51) Ethanol (9) None (52) Ethanol (9)

Number of replicates (Q4) 2 (52) 4 (10) 2 (56) 4 (10)

Approach to test for PCR inhibition (Q5)

OD ratios (40) delta Cq or GM between two dilutions (31)

OD ratios (36) delta Cq or GM between two dilutions (32)

Reason for not testing all events (Q9)

Not applicable (45) below the LOQ (18) Not applicable (61) reagents not available (10)

Approach used (Q6a) Standard curves (68) delta Cq (9) Standard curves (71) delta Cq (9)

Calibrant used (Q6b) CRM JRC-Geel in gkg (71) other RM in copies (4) CRM JRC-Geel in gkg (74) other RM in copies (4)

Taxon-specific endogenous gene (Q6c)

lec-74 bp (65) lec-81 or 118 bp (each 5) lec-74 bp (63) lec-118 bp (6)

Unit of measurement and data expression (Q6d)

Mass (59) copies=mass CRM (13) Mass (61) copies=mass CRM (14)

Amount of DNA (Q6e) 200 ng (30) 100 ng (18) 200 ng (30) 100 ng (21)

LOQ (Q6f) 01 (37) lt01 (34) 01 (39) lt01 (39)

LOQ determination (Q6g) In-house validation (42) current analysis (20) In-house validation (43) current analysis (21)

Uncertainty determination (Q6h)

Precision of replicates (32) in-house validation (27) Precision of replicates (32) in-house validation (30)

1 The numbers shown refer to the number of laboratories that reported the answer The answers that were reported with the two largest frequencies are mentioned

One NRL882 (L73) reported that T1 was out of the scope of the laboratory and one non-

NRL (L43) reported that the T1 matrix was not analysed because of practical constraints

The evaluation of the answers shows that the most commonly employed DNA extraction

method for both T1 and T2 was one based on CTAB with the NucleoSpin Food kit ranking

second No additional purification methods were generally applied The majority of

laboratories analysed two replicate DNA extracts Most laboratories checked the quality of

the DNA extracts by verifying the OD ratios andor running two dilutions a minority of

laboratories performed a PCR inhibition run on 3 or 4 DNA dilutions with a reference gene

For the quantitative analysis the most common approach used was based on two standard

curves however 9 laboratories applied the delta Cq approach One laboratory (L50)

mentioned the use of digital PCR for 40-3-2 soybean quantification in T1 and T2 The

available CRMs from the JRC were used by most laboratories but 5 laboratories used a

non-certified reference material (RM) where values were expressed in GM copy number

ratio (4) or GM mass fraction (1) Lec was used as taxon-specific reference gene by all

laboratories for soybean (mostly the 74 bp version) The majority of laboratories

performed their measurements in the same unit as the certified value of the calibrant used

(gkg) and no conversion factor was applied The LOQ reported was either taken from in-

house validation of the method or determined from the analysis results for this CT In

most cases a LOQ of 01 mm or lower was reported The measurement uncertainty

was either derived from the standard deviation of the measurement replicates or from the

intermediate precision determined in the frame of the single-laboratory validation study

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

846

2 Test items

The test items used in this PT were prepared and characterised at JRC-Geel

21 Test item 1

The T1 test item was prepared from base materials that were characterised before their

use (Table 1) The base materials employed for the preparation of T1 were chicken feed

(AVEVE for biological agriculture according to EC 8342007 and EC 8892008) the ERM-

BF410ak as non GM soybean powder and the ERM-BF410bp(6) containing the MON-Oslash4Oslash32-

6 event (hereafter named 40-3-2) as spiking material (Table 1)

The chicken feed was composed according to the label of Bio maize kernels Bio soybean

oil-cake Bio-sunflower oil-cake Bio wheat Bio barley maize gluten potato proteins

calcium monophosphate calcium carbonate Bio soybean oil and sodium chloride The

analytical composition indicated a content of 175 protein 5 fat 13 ashes 5

cellulose 033 methionine 078 lysine 37 calcium 052 phosphorus and

01 sodium The chicken feed was milled using a cryo-grinding vibrating mill (Palla mill

KHD Humboldt-Wedag Koumlln DE) and sieved with a 500 microm stainless steel mesh on a

sieving machine equipped with an ultrasonic sieving aid (Russel Finex London UK) The

remaining powder was mixed in a DynaMIX CM200 (WAB Muttenz CH) for 1 h to

homogenise the distribution of the different types of seed tissues

The residual water mass fractions for the chicken feed powder and the powders of the

certified reference materials ERM-BF410ak and ERM-BF410bp were measured by

volumetric Karl Fischer titration (758 KFD Titrino Metrohm Herisau CH) The results

showed that the powders were sufficiently dry to perform the dry mixing and did not

require an additional drying step

The particle size distribution of the powders was measured using laser diffraction (PSA

Sympatec Clausthal-Zellerfeld DE) It was concluded that the particle size distribution of

these powders was sufficiently similar to allow subsequent preparations of mixtures

The amount and the quality of the DNA extracted from the chicken feed powder the non-

GM soybean flour and the GM spiking material were verified by UV spectrometry

fluorometry and gel electrophoresis A CTAB-tip 20G method (Qiagen Hilden Germany)

optimised for soybean was chosen with a sample intake of 200 mg because it yielded a

sufficient amount of DNA of PCR-grade quality from the base materials DNA extracted

with the in-house CTAB method was tested for PCR inhibition between 40 ngmicroL to 02

ngmicroL with a lectin qPCR assay (5 microL per PCR) and did not show any inhibition (Cq values

were very close to the theoretical Cq values) The PCR efficiencies ranged from 96 to 98

with a coefficient of determination (R2) between 099 and 100 confirming the absence

of significant amounts of PCR inhibitors in the extracts

The CTAB method yielded a sufficient amount of DNA of PCR-grade quality from both non-

GM and GM base materials

The level of fragmentation of the extracted DNA was investigated by 10 agarose gel

electrophoresis A smear from plusmn 12 to 1 kbp could be clearly seen in the chicken feed

DNA indicating some level of fragmentation of the extracted DNA while the DNA

extracted from the soybean materials migrated as a high molecular weight band (above 12

kbp) The amount of soybean DNA that could be extracted and amplified from the chicken

feed powder was determined by qPCR with a lectin assay using DNA from a soybean CRM

as calibrant this amount appeared to be rather low (lt1 of total DNA) The yield of

amplifiable DNA per mg of chicken feed powder and the yield of DNA measured by

PicoGreen for the soybean materials composed of pure soybean were taken into account

to calculate the amount of chicken feed powder non-GM soybean and GM soybean to be

mixed to obtain a target value of approximately 08 mm event 40-3-2 in T1

EURL GMFF CT 0217 report

946

Table 1 Characteristics of the base materials used for preparation of test item 1 (T1)

Characteristic Chicken Feed Non-GM

Soybean Flour

40-3-2

Soybean Flour

Type of base material Scratch grains CRM CRM

Origin AVEVE (Belgium) ERM-BF410ak ERM-BF410bp

Grinding method Cryo-grinding vibrating mill Used as such Used as such

Mixing method DynaMIX CM200 (WAB Muttenz CH)

Water content in gkg mean plusmn U

(k = 2 n = 3) 106 plusmn 15 115 plusmn 07 155 plusmn 10

Particle diameter in microm mean plusmn U

(k = 2 n = 5) 1114 plusmn 198 1188 plusmn 272 1213 plusmn 60

Soybean DNA yield in ngmg1 mean plusmn U

(k = 2) 07 plusmn 01 (n=5) 579 plusmn 178 (n=3) 729 plusmn 18 (n=2)

Genetic elements detected with screening

pre-spotted plates (Cq value)2

Hmg (Cq 246) Lec (247) CruA (Cq 357) UGP (281) PLD (365) P35S (Cq 333) tNOS (Cq 368) CTP2-EPSPS (Cq 357) Cry1AbAc (Cq 360) PAT (Cq 376)

Lec (Cq 215) Lec (Cq 212) p35S (Cq 212) tNOS (Cq 225) PAT (Cq 348)

GM soybean events detected with event-

specific pre-spotted plates (Cq value)2

40-3-2 (Cq 340) MON87701 (Cq 374) MON89788 (Cq 363) Lec (Cq 246)

NA 40-3-2 (Cq 214) A5547 (Cq 412)

Mass used to prepare T1 (g) 11032 96 108

Nominal target GM mass fraction in T1

(mm ) NA NA 08

1 Results reported here for a sample intake of 200 mg with the in-house validated CTAB method + Genomic-tip 20G purification for soybean (JRC-GEEL) The soybean DNA yield value for the chicken feed was determined by qPCR whereas the yield from the non-GM and GM soybean materials was measured by fluorometry 2 A screening and GM soybean event-specific pre-spotted plate (PSP) was used for these tests NA not applicable k coverage factor U expanded uncertainty

The presence of different species and GM events in the base materials and in a pilot

mixture was tested by using the screening(7) and GM soybean event-specific pre-spotted

plates(8)

The presence of maize and soybean in the chicken feed powder was confirmed by the early

quantification cycle obtained for the high mobility gene (hmg) and lectin (lec) assays Late

amplifications for the UDP-glucose pyrophosphorylase (ugp) phospholipase D (PLD) and

cruciferin A (cruA) genes confirmed the respective presence of potato sunflower and

rapeseed in the chicken feed

The chicken feed powder (labelled bio) also contained traces of genetic markers such as

p35S tNOS EPSPS and Cry1AbAc which indicate a contamination of the chicken feed by

genetically modified plant materials The GM soybean event-specific assays confirmed the

presence of traces of 3 GM soybeans namely 40-3-2 MON87701 and MON89788 The

level of contamination was estimated to be below 001

The final test item was prepared gravimetrically in accordance with ISO 170342016(9) as

follows

The mass of the GM ingredient to add (40-3-2 soybean) was calculated taking into

account the amount of DNA that could be extracted and amplified from the different

materials (Table 1)

The compound sample T1 was mixed in a DynaMIX CM200 for 15 h to improve

homogeneity

After finalisation of the mixing step the powders were filled manually in 20 mL

brown glass vials using lyophilisation inserts manually placed in the bottle necks

Before final closure of the vials air was evacuated in a freeze-dryer and replaced

by argon The vials were finally closed inside the freeze-dryer with the help of a

hydraulic device and then sealed with blue aluminium caps to maintain the inert

atmosphere and to prevent accidental opening during storage and transport

A total of 200 vials containing each at least 5 g of flour were then labelled with a

sample number and the description Sample T1 (chicken feed)

EURL GMFF CT 0217 report

1046

Following the inventory and the selection of vials for future analysis according to a

random stratified sampling scheme the bottles were brought to a storage room for

long-term storage in the dark at 4 plusmn 3 degC

Homogeneity and stability testing of T1 was performed in-house as described in Annex 1

using an event-specific quantification method previously validated by the EURL GMFF

Material T1 was found to be homogeneous for the GM event added (p-value gt 005)

based on a 200 mg sample intake

From the isochronous stability study it was concluded that the test item would be

sufficiently stable under ambient shipment conditions (5 significance level) Stability

was also confirmed during the whole period of the PT between the dispatch of the test

items until the deadline of reporting the results (Annex 1)

JRC-Ispra tested the T1 material and this confirmed the results obtained by JRC-Geel The

average (n = 91) mass fraction of event 40-3-2 measured in T1 was 069 plusmn 004 mm

(U k = 2) which approximated the expected nominal value but may have been

influenced by the characteristics of the different base materials

22 Test item 2

The T2 test item was a new batch of a certified reference material that was not yet

released on the market (Table 2) The bottles of T2 were re-labelled with a unique sample

number as Sample T2 (soybean flour)

Homogeneity and short-term stability of T2 had been previously demonstrated as part of

the certification of the CRM stability monitoring confirmed the stability of T2 during the

running time of the PT (Annex 1)

Table 2 Characteristics of test item 2 (T2)

Characteristic Soybean feed

Type of base material CRM

Origin ERM-BF410dp(6) containing 100 plusmn 06 gkg MON-Oslash4Oslash32-6 soybean produced in 2017 by JRC-Geel

The certificate of ERM-BF410dp warns that a difference (at 95 confidence level)

between the total DNA content in the two powders used for the production of ERM-

BF410dp was found to be significant (due to the different size of non-GM and GM seeds)

and is likely to have an impact when using this CRM Depending on the composition of the

unknown sample real-time PCR measurement results of ERM-BF410dp may differ up to

239 plusmn 11 (average plusmn U) compared to the results of the unknown sample This

difference may depend also on the DNA extraction method selected and both effects may

be additive This observation was confirmed by the participants in this PT who used the

previous 40-3-2 soybean CRM batch (ERM-BF410n) for the calibration of the T2

measurements

EURL GMFF CT 0217 report

1146

3 Instructions to the participants

Participants in this PT were instructed to analyse the two test items (T1 and T2) as

follows

Test Item 1 Chicken feed

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Test Item 2 Soybean flour

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Quantitative results had to be reported in mm as outlined below

Mass GM event [g]

mm = x 100 (1)

Total mass species [g]

Participants were requested

- to use procedures for detectionquantification of the GM events that resemble as

closely as possible the ones used in routine sample analysis

- to take care in ensuring that the DNA extraction method employed is adapted to

the matrix and that the quality of the DNA obtained is suitable for PCR

- to report the quantitative results with two decimal places (eg 064 or 129)

- to follow the general rule that results obtained using a calibrant certified for GM

mass fraction (ie a matrix CRM certified in [x] gkg) can directly be expressed in

mm while results obtained using a calibrant certified for DNA copy number

ratio (eg a plasmid containing both the GM and reference gene target or some

matrix CRMs) need to be converted into mm using a conversion factor(1011)

- to pay attention to the correct estimation and reporting of the measurement

uncertainty and coverage factor used as the uncertainty reported would be

considered in the evaluation of the results using zeta scores

- in case of an unsatisfactory performance to fill in a form indicating the root-cause

analysis and providing evidence demonstrating the effectiveness of the corrective

actions implemented (for NRLs only)

The participants were also informed that the identification information on the participants

in this PT would be kept confidential except for the NRLs that have been appointed in line

with Regulation (EC) No 8822004 their lab codes will be disclosed to DG SANTE for the

purpose of an assessment of their performance

EURL GMFF CT 0217 report

1246

4 Results

41 Participation to the PT

On 31 May 2017 199 laboratories were informed about the upcoming PT EURL-GMFF-CT-

0217 Finally 89 laboratories registered for it and received a random unique lab code

(L01 to L89) Eighty-six laboratories from 39 countries returned results within the

reporting deadline Three non-NRLs did not submit any results two of which (L21 L28)

had not received the samples from customs while L67 did not provide any justification for

not participating

Table 3 shows an overview of the participation to this PT

Table 3 Communication about and participation to the PT 0217

Characteristic of the PT Result

Date of PT announcement 31 May 2017

Deadline for registration 14 June 2017

Date of shipment of samples 3 July 2017

Deadline for result submission 25 August 2017

Number of laboratories informed 199

Number of registered laboratories 89

Registered laboratories that failed to submit their data 3

Number of participating laboratories 86

The participating laboratories fell into the following assigned categories (Table 4)

Thirty-three NRLs designated in line with Regulation (EC) No 8822004 (NRL882)

representing 25 EU Member States (many of them are also NRL120) In addition

Ireland delegated its NRL882 tasks to one of the PT participants Estonia and

Malta were not represented in this PT

Twenty-one NRLs nominated under Regulation (EU) No 1202014 (NRL120) that

are not at the same time NRL under Regulation (EC) No 8822004

Thirty-two laboratories that are not NRL but are appointed by their National

Authority to perform official controls This category includes 10 EU official control

laboratories (OCLs) and 22 laboratories from non-EU countries including Serbia

and Switzerland

Among the countries Germany was represented with 17 laboratories Italy with 6

laboratories and Belgium and Poland with 4 laboratories each all other countries had

between one and three participating laboratories

EURL GMFF CT 0217 report

1346

Table 4 Overview of participants to CT 0217 by country and category

Country Participants NRL8821 NRL120 Non-NRL

AUSTRIA 2 2

BELGIUM 4 3

1

BRAZIL 2

2

BULGARIA 2 1

1

CHILE 1

1

COLOMBIA 1

1

CROATIA 2 1

1

CYPRUS 1 1

CZECH REPUBLIC 1 1

DENMARK 1 1

FINLAND 2 1 1

FRANCE 2 2

GERMANY 17 1 14 2

GREECE 1 1

HUNGARY 2 1

1

INDIA 1

1

ITALY 6 1 2 3

LATVIA 1 1

LEBANON 1

1

LITHUANIA 1 1

LUXEMBOURG 1 1

MEXICO 1

1

NETHERLANDS 2 1 1

PHILIPPINES 1

1

POLAND 4 3 1

PORTUGAL 1 1

ROMANIA 2 1

1

SERBIA 3

3

SINGAPORE 1

1

SLOVAKIA 2 2

SLOVENIA 1 1

SPAIN 2 2

SWEDEN 1 1

SWITZERLAND 2

2

TURKEY 1

1

UKRAINE 3

3

UNITED KINGDOM 3 1 2

UNITED STATES 1

1

VIETNAM 3

3

Total 86 33 21 32 1 No NRL882 from Estonia or Malta participated to this PT

EURL GMFF CT 0217 report

1446

42 Information on the testing provided in the questionnaire

Participants were asked to fill in an online questionnaire (through EUSurvey) on their

testing methodology used for T1 and T2 consisting of a number of mostly multiple-choice

questions A total of 82 laboratories completed the questionnaire including all 33

NRL882 20 out of 21 NRL120 and 29 out of 32 non-NRLs (questionnaires were missing

from L03 L10 L38 and L49)

Table 5 summarises the main answers received whereas Annex 2 shows all answers The

results on GM event identification are reported in Section 43

Table 5 Summary of the main answers provided in the questionnaire of CT 0217

Question (and Question

number) Test Item 1 ndash 40-3-2 Test Item 2 ndash 40-3-2

Test item analysed Yes (801) No (2) Yes (82) No (0)

Reason for lack of analysis (Q1)

Matrix out of scope (1) other practical constraints (1) -

DNA extraction method (Q2) CTAB (41) NucleoSpin Food (10) CTAB (40) NucleoSpin Food (11)

Additional DNA purification method (Q3)

None (51) Ethanol (9) None (52) Ethanol (9)

Number of replicates (Q4) 2 (52) 4 (10) 2 (56) 4 (10)

Approach to test for PCR inhibition (Q5)

OD ratios (40) delta Cq or GM between two dilutions (31)

OD ratios (36) delta Cq or GM between two dilutions (32)

Reason for not testing all events (Q9)

Not applicable (45) below the LOQ (18) Not applicable (61) reagents not available (10)

Approach used (Q6a) Standard curves (68) delta Cq (9) Standard curves (71) delta Cq (9)

Calibrant used (Q6b) CRM JRC-Geel in gkg (71) other RM in copies (4) CRM JRC-Geel in gkg (74) other RM in copies (4)

Taxon-specific endogenous gene (Q6c)

lec-74 bp (65) lec-81 or 118 bp (each 5) lec-74 bp (63) lec-118 bp (6)

Unit of measurement and data expression (Q6d)

Mass (59) copies=mass CRM (13) Mass (61) copies=mass CRM (14)

Amount of DNA (Q6e) 200 ng (30) 100 ng (18) 200 ng (30) 100 ng (21)

LOQ (Q6f) 01 (37) lt01 (34) 01 (39) lt01 (39)

LOQ determination (Q6g) In-house validation (42) current analysis (20) In-house validation (43) current analysis (21)

Uncertainty determination (Q6h)

Precision of replicates (32) in-house validation (27) Precision of replicates (32) in-house validation (30)

1 The numbers shown refer to the number of laboratories that reported the answer The answers that were reported with the two largest frequencies are mentioned

One NRL882 (L73) reported that T1 was out of the scope of the laboratory and one non-

NRL (L43) reported that the T1 matrix was not analysed because of practical constraints

The evaluation of the answers shows that the most commonly employed DNA extraction

method for both T1 and T2 was one based on CTAB with the NucleoSpin Food kit ranking

second No additional purification methods were generally applied The majority of

laboratories analysed two replicate DNA extracts Most laboratories checked the quality of

the DNA extracts by verifying the OD ratios andor running two dilutions a minority of

laboratories performed a PCR inhibition run on 3 or 4 DNA dilutions with a reference gene

For the quantitative analysis the most common approach used was based on two standard

curves however 9 laboratories applied the delta Cq approach One laboratory (L50)

mentioned the use of digital PCR for 40-3-2 soybean quantification in T1 and T2 The

available CRMs from the JRC were used by most laboratories but 5 laboratories used a

non-certified reference material (RM) where values were expressed in GM copy number

ratio (4) or GM mass fraction (1) Lec was used as taxon-specific reference gene by all

laboratories for soybean (mostly the 74 bp version) The majority of laboratories

performed their measurements in the same unit as the certified value of the calibrant used

(gkg) and no conversion factor was applied The LOQ reported was either taken from in-

house validation of the method or determined from the analysis results for this CT In

most cases a LOQ of 01 mm or lower was reported The measurement uncertainty

was either derived from the standard deviation of the measurement replicates or from the

intermediate precision determined in the frame of the single-laboratory validation study

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

946

Table 1 Characteristics of the base materials used for preparation of test item 1 (T1)

Characteristic Chicken Feed Non-GM

Soybean Flour

40-3-2

Soybean Flour

Type of base material Scratch grains CRM CRM

Origin AVEVE (Belgium) ERM-BF410ak ERM-BF410bp

Grinding method Cryo-grinding vibrating mill Used as such Used as such

Mixing method DynaMIX CM200 (WAB Muttenz CH)

Water content in gkg mean plusmn U

(k = 2 n = 3) 106 plusmn 15 115 plusmn 07 155 plusmn 10

Particle diameter in microm mean plusmn U

(k = 2 n = 5) 1114 plusmn 198 1188 plusmn 272 1213 plusmn 60

Soybean DNA yield in ngmg1 mean plusmn U

(k = 2) 07 plusmn 01 (n=5) 579 plusmn 178 (n=3) 729 plusmn 18 (n=2)

Genetic elements detected with screening

pre-spotted plates (Cq value)2

Hmg (Cq 246) Lec (247) CruA (Cq 357) UGP (281) PLD (365) P35S (Cq 333) tNOS (Cq 368) CTP2-EPSPS (Cq 357) Cry1AbAc (Cq 360) PAT (Cq 376)

Lec (Cq 215) Lec (Cq 212) p35S (Cq 212) tNOS (Cq 225) PAT (Cq 348)

GM soybean events detected with event-

specific pre-spotted plates (Cq value)2

40-3-2 (Cq 340) MON87701 (Cq 374) MON89788 (Cq 363) Lec (Cq 246)

NA 40-3-2 (Cq 214) A5547 (Cq 412)

Mass used to prepare T1 (g) 11032 96 108

Nominal target GM mass fraction in T1

(mm ) NA NA 08

1 Results reported here for a sample intake of 200 mg with the in-house validated CTAB method + Genomic-tip 20G purification for soybean (JRC-GEEL) The soybean DNA yield value for the chicken feed was determined by qPCR whereas the yield from the non-GM and GM soybean materials was measured by fluorometry 2 A screening and GM soybean event-specific pre-spotted plate (PSP) was used for these tests NA not applicable k coverage factor U expanded uncertainty

The presence of different species and GM events in the base materials and in a pilot

mixture was tested by using the screening(7) and GM soybean event-specific pre-spotted

plates(8)

The presence of maize and soybean in the chicken feed powder was confirmed by the early

quantification cycle obtained for the high mobility gene (hmg) and lectin (lec) assays Late

amplifications for the UDP-glucose pyrophosphorylase (ugp) phospholipase D (PLD) and

cruciferin A (cruA) genes confirmed the respective presence of potato sunflower and

rapeseed in the chicken feed

The chicken feed powder (labelled bio) also contained traces of genetic markers such as

p35S tNOS EPSPS and Cry1AbAc which indicate a contamination of the chicken feed by

genetically modified plant materials The GM soybean event-specific assays confirmed the

presence of traces of 3 GM soybeans namely 40-3-2 MON87701 and MON89788 The

level of contamination was estimated to be below 001

The final test item was prepared gravimetrically in accordance with ISO 170342016(9) as

follows

The mass of the GM ingredient to add (40-3-2 soybean) was calculated taking into

account the amount of DNA that could be extracted and amplified from the different

materials (Table 1)

The compound sample T1 was mixed in a DynaMIX CM200 for 15 h to improve

homogeneity

After finalisation of the mixing step the powders were filled manually in 20 mL

brown glass vials using lyophilisation inserts manually placed in the bottle necks

Before final closure of the vials air was evacuated in a freeze-dryer and replaced

by argon The vials were finally closed inside the freeze-dryer with the help of a

hydraulic device and then sealed with blue aluminium caps to maintain the inert

atmosphere and to prevent accidental opening during storage and transport

A total of 200 vials containing each at least 5 g of flour were then labelled with a

sample number and the description Sample T1 (chicken feed)

EURL GMFF CT 0217 report

1046

Following the inventory and the selection of vials for future analysis according to a

random stratified sampling scheme the bottles were brought to a storage room for

long-term storage in the dark at 4 plusmn 3 degC

Homogeneity and stability testing of T1 was performed in-house as described in Annex 1

using an event-specific quantification method previously validated by the EURL GMFF

Material T1 was found to be homogeneous for the GM event added (p-value gt 005)

based on a 200 mg sample intake

From the isochronous stability study it was concluded that the test item would be

sufficiently stable under ambient shipment conditions (5 significance level) Stability

was also confirmed during the whole period of the PT between the dispatch of the test

items until the deadline of reporting the results (Annex 1)

JRC-Ispra tested the T1 material and this confirmed the results obtained by JRC-Geel The

average (n = 91) mass fraction of event 40-3-2 measured in T1 was 069 plusmn 004 mm

(U k = 2) which approximated the expected nominal value but may have been

influenced by the characteristics of the different base materials

22 Test item 2

The T2 test item was a new batch of a certified reference material that was not yet

released on the market (Table 2) The bottles of T2 were re-labelled with a unique sample

number as Sample T2 (soybean flour)

Homogeneity and short-term stability of T2 had been previously demonstrated as part of

the certification of the CRM stability monitoring confirmed the stability of T2 during the

running time of the PT (Annex 1)

Table 2 Characteristics of test item 2 (T2)

Characteristic Soybean feed

Type of base material CRM

Origin ERM-BF410dp(6) containing 100 plusmn 06 gkg MON-Oslash4Oslash32-6 soybean produced in 2017 by JRC-Geel

The certificate of ERM-BF410dp warns that a difference (at 95 confidence level)

between the total DNA content in the two powders used for the production of ERM-

BF410dp was found to be significant (due to the different size of non-GM and GM seeds)

and is likely to have an impact when using this CRM Depending on the composition of the

unknown sample real-time PCR measurement results of ERM-BF410dp may differ up to

239 plusmn 11 (average plusmn U) compared to the results of the unknown sample This

difference may depend also on the DNA extraction method selected and both effects may

be additive This observation was confirmed by the participants in this PT who used the

previous 40-3-2 soybean CRM batch (ERM-BF410n) for the calibration of the T2

measurements

EURL GMFF CT 0217 report

1146

3 Instructions to the participants

Participants in this PT were instructed to analyse the two test items (T1 and T2) as

follows

Test Item 1 Chicken feed

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Test Item 2 Soybean flour

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Quantitative results had to be reported in mm as outlined below

Mass GM event [g]

mm = x 100 (1)

Total mass species [g]

Participants were requested

- to use procedures for detectionquantification of the GM events that resemble as

closely as possible the ones used in routine sample analysis

- to take care in ensuring that the DNA extraction method employed is adapted to

the matrix and that the quality of the DNA obtained is suitable for PCR

- to report the quantitative results with two decimal places (eg 064 or 129)

- to follow the general rule that results obtained using a calibrant certified for GM

mass fraction (ie a matrix CRM certified in [x] gkg) can directly be expressed in

mm while results obtained using a calibrant certified for DNA copy number

ratio (eg a plasmid containing both the GM and reference gene target or some

matrix CRMs) need to be converted into mm using a conversion factor(1011)

- to pay attention to the correct estimation and reporting of the measurement

uncertainty and coverage factor used as the uncertainty reported would be

considered in the evaluation of the results using zeta scores

- in case of an unsatisfactory performance to fill in a form indicating the root-cause

analysis and providing evidence demonstrating the effectiveness of the corrective

actions implemented (for NRLs only)

The participants were also informed that the identification information on the participants

in this PT would be kept confidential except for the NRLs that have been appointed in line

with Regulation (EC) No 8822004 their lab codes will be disclosed to DG SANTE for the

purpose of an assessment of their performance

EURL GMFF CT 0217 report

1246

4 Results

41 Participation to the PT

On 31 May 2017 199 laboratories were informed about the upcoming PT EURL-GMFF-CT-

0217 Finally 89 laboratories registered for it and received a random unique lab code

(L01 to L89) Eighty-six laboratories from 39 countries returned results within the

reporting deadline Three non-NRLs did not submit any results two of which (L21 L28)

had not received the samples from customs while L67 did not provide any justification for

not participating

Table 3 shows an overview of the participation to this PT

Table 3 Communication about and participation to the PT 0217

Characteristic of the PT Result

Date of PT announcement 31 May 2017

Deadline for registration 14 June 2017

Date of shipment of samples 3 July 2017

Deadline for result submission 25 August 2017

Number of laboratories informed 199

Number of registered laboratories 89

Registered laboratories that failed to submit their data 3

Number of participating laboratories 86

The participating laboratories fell into the following assigned categories (Table 4)

Thirty-three NRLs designated in line with Regulation (EC) No 8822004 (NRL882)

representing 25 EU Member States (many of them are also NRL120) In addition

Ireland delegated its NRL882 tasks to one of the PT participants Estonia and

Malta were not represented in this PT

Twenty-one NRLs nominated under Regulation (EU) No 1202014 (NRL120) that

are not at the same time NRL under Regulation (EC) No 8822004

Thirty-two laboratories that are not NRL but are appointed by their National

Authority to perform official controls This category includes 10 EU official control

laboratories (OCLs) and 22 laboratories from non-EU countries including Serbia

and Switzerland

Among the countries Germany was represented with 17 laboratories Italy with 6

laboratories and Belgium and Poland with 4 laboratories each all other countries had

between one and three participating laboratories

EURL GMFF CT 0217 report

1346

Table 4 Overview of participants to CT 0217 by country and category

Country Participants NRL8821 NRL120 Non-NRL

AUSTRIA 2 2

BELGIUM 4 3

1

BRAZIL 2

2

BULGARIA 2 1

1

CHILE 1

1

COLOMBIA 1

1

CROATIA 2 1

1

CYPRUS 1 1

CZECH REPUBLIC 1 1

DENMARK 1 1

FINLAND 2 1 1

FRANCE 2 2

GERMANY 17 1 14 2

GREECE 1 1

HUNGARY 2 1

1

INDIA 1

1

ITALY 6 1 2 3

LATVIA 1 1

LEBANON 1

1

LITHUANIA 1 1

LUXEMBOURG 1 1

MEXICO 1

1

NETHERLANDS 2 1 1

PHILIPPINES 1

1

POLAND 4 3 1

PORTUGAL 1 1

ROMANIA 2 1

1

SERBIA 3

3

SINGAPORE 1

1

SLOVAKIA 2 2

SLOVENIA 1 1

SPAIN 2 2

SWEDEN 1 1

SWITZERLAND 2

2

TURKEY 1

1

UKRAINE 3

3

UNITED KINGDOM 3 1 2

UNITED STATES 1

1

VIETNAM 3

3

Total 86 33 21 32 1 No NRL882 from Estonia or Malta participated to this PT

EURL GMFF CT 0217 report

1446

42 Information on the testing provided in the questionnaire

Participants were asked to fill in an online questionnaire (through EUSurvey) on their

testing methodology used for T1 and T2 consisting of a number of mostly multiple-choice

questions A total of 82 laboratories completed the questionnaire including all 33

NRL882 20 out of 21 NRL120 and 29 out of 32 non-NRLs (questionnaires were missing

from L03 L10 L38 and L49)

Table 5 summarises the main answers received whereas Annex 2 shows all answers The

results on GM event identification are reported in Section 43

Table 5 Summary of the main answers provided in the questionnaire of CT 0217

Question (and Question

number) Test Item 1 ndash 40-3-2 Test Item 2 ndash 40-3-2

Test item analysed Yes (801) No (2) Yes (82) No (0)

Reason for lack of analysis (Q1)

Matrix out of scope (1) other practical constraints (1) -

DNA extraction method (Q2) CTAB (41) NucleoSpin Food (10) CTAB (40) NucleoSpin Food (11)

Additional DNA purification method (Q3)

None (51) Ethanol (9) None (52) Ethanol (9)

Number of replicates (Q4) 2 (52) 4 (10) 2 (56) 4 (10)

Approach to test for PCR inhibition (Q5)

OD ratios (40) delta Cq or GM between two dilutions (31)

OD ratios (36) delta Cq or GM between two dilutions (32)

Reason for not testing all events (Q9)

Not applicable (45) below the LOQ (18) Not applicable (61) reagents not available (10)

Approach used (Q6a) Standard curves (68) delta Cq (9) Standard curves (71) delta Cq (9)

Calibrant used (Q6b) CRM JRC-Geel in gkg (71) other RM in copies (4) CRM JRC-Geel in gkg (74) other RM in copies (4)

Taxon-specific endogenous gene (Q6c)

lec-74 bp (65) lec-81 or 118 bp (each 5) lec-74 bp (63) lec-118 bp (6)

Unit of measurement and data expression (Q6d)

Mass (59) copies=mass CRM (13) Mass (61) copies=mass CRM (14)

Amount of DNA (Q6e) 200 ng (30) 100 ng (18) 200 ng (30) 100 ng (21)

LOQ (Q6f) 01 (37) lt01 (34) 01 (39) lt01 (39)

LOQ determination (Q6g) In-house validation (42) current analysis (20) In-house validation (43) current analysis (21)

Uncertainty determination (Q6h)

Precision of replicates (32) in-house validation (27) Precision of replicates (32) in-house validation (30)

1 The numbers shown refer to the number of laboratories that reported the answer The answers that were reported with the two largest frequencies are mentioned

One NRL882 (L73) reported that T1 was out of the scope of the laboratory and one non-

NRL (L43) reported that the T1 matrix was not analysed because of practical constraints

The evaluation of the answers shows that the most commonly employed DNA extraction

method for both T1 and T2 was one based on CTAB with the NucleoSpin Food kit ranking

second No additional purification methods were generally applied The majority of

laboratories analysed two replicate DNA extracts Most laboratories checked the quality of

the DNA extracts by verifying the OD ratios andor running two dilutions a minority of

laboratories performed a PCR inhibition run on 3 or 4 DNA dilutions with a reference gene

For the quantitative analysis the most common approach used was based on two standard

curves however 9 laboratories applied the delta Cq approach One laboratory (L50)

mentioned the use of digital PCR for 40-3-2 soybean quantification in T1 and T2 The

available CRMs from the JRC were used by most laboratories but 5 laboratories used a

non-certified reference material (RM) where values were expressed in GM copy number

ratio (4) or GM mass fraction (1) Lec was used as taxon-specific reference gene by all

laboratories for soybean (mostly the 74 bp version) The majority of laboratories

performed their measurements in the same unit as the certified value of the calibrant used

(gkg) and no conversion factor was applied The LOQ reported was either taken from in-

house validation of the method or determined from the analysis results for this CT In

most cases a LOQ of 01 mm or lower was reported The measurement uncertainty

was either derived from the standard deviation of the measurement replicates or from the

intermediate precision determined in the frame of the single-laboratory validation study

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

1046

Following the inventory and the selection of vials for future analysis according to a

random stratified sampling scheme the bottles were brought to a storage room for

long-term storage in the dark at 4 plusmn 3 degC

Homogeneity and stability testing of T1 was performed in-house as described in Annex 1

using an event-specific quantification method previously validated by the EURL GMFF

Material T1 was found to be homogeneous for the GM event added (p-value gt 005)

based on a 200 mg sample intake

From the isochronous stability study it was concluded that the test item would be

sufficiently stable under ambient shipment conditions (5 significance level) Stability

was also confirmed during the whole period of the PT between the dispatch of the test

items until the deadline of reporting the results (Annex 1)

JRC-Ispra tested the T1 material and this confirmed the results obtained by JRC-Geel The

average (n = 91) mass fraction of event 40-3-2 measured in T1 was 069 plusmn 004 mm

(U k = 2) which approximated the expected nominal value but may have been

influenced by the characteristics of the different base materials

22 Test item 2

The T2 test item was a new batch of a certified reference material that was not yet

released on the market (Table 2) The bottles of T2 were re-labelled with a unique sample

number as Sample T2 (soybean flour)

Homogeneity and short-term stability of T2 had been previously demonstrated as part of

the certification of the CRM stability monitoring confirmed the stability of T2 during the

running time of the PT (Annex 1)

Table 2 Characteristics of test item 2 (T2)

Characteristic Soybean feed

Type of base material CRM

Origin ERM-BF410dp(6) containing 100 plusmn 06 gkg MON-Oslash4Oslash32-6 soybean produced in 2017 by JRC-Geel

The certificate of ERM-BF410dp warns that a difference (at 95 confidence level)

between the total DNA content in the two powders used for the production of ERM-

BF410dp was found to be significant (due to the different size of non-GM and GM seeds)

and is likely to have an impact when using this CRM Depending on the composition of the

unknown sample real-time PCR measurement results of ERM-BF410dp may differ up to

239 plusmn 11 (average plusmn U) compared to the results of the unknown sample This

difference may depend also on the DNA extraction method selected and both effects may

be additive This observation was confirmed by the participants in this PT who used the

previous 40-3-2 soybean CRM batch (ERM-BF410n) for the calibration of the T2

measurements

EURL GMFF CT 0217 report

1146

3 Instructions to the participants

Participants in this PT were instructed to analyse the two test items (T1 and T2) as

follows

Test Item 1 Chicken feed

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Test Item 2 Soybean flour

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Quantitative results had to be reported in mm as outlined below

Mass GM event [g]

mm = x 100 (1)

Total mass species [g]

Participants were requested

- to use procedures for detectionquantification of the GM events that resemble as

closely as possible the ones used in routine sample analysis

- to take care in ensuring that the DNA extraction method employed is adapted to

the matrix and that the quality of the DNA obtained is suitable for PCR

- to report the quantitative results with two decimal places (eg 064 or 129)

- to follow the general rule that results obtained using a calibrant certified for GM

mass fraction (ie a matrix CRM certified in [x] gkg) can directly be expressed in

mm while results obtained using a calibrant certified for DNA copy number

ratio (eg a plasmid containing both the GM and reference gene target or some

matrix CRMs) need to be converted into mm using a conversion factor(1011)

- to pay attention to the correct estimation and reporting of the measurement

uncertainty and coverage factor used as the uncertainty reported would be

considered in the evaluation of the results using zeta scores

- in case of an unsatisfactory performance to fill in a form indicating the root-cause

analysis and providing evidence demonstrating the effectiveness of the corrective

actions implemented (for NRLs only)

The participants were also informed that the identification information on the participants

in this PT would be kept confidential except for the NRLs that have been appointed in line

with Regulation (EC) No 8822004 their lab codes will be disclosed to DG SANTE for the

purpose of an assessment of their performance

EURL GMFF CT 0217 report

1246

4 Results

41 Participation to the PT

On 31 May 2017 199 laboratories were informed about the upcoming PT EURL-GMFF-CT-

0217 Finally 89 laboratories registered for it and received a random unique lab code

(L01 to L89) Eighty-six laboratories from 39 countries returned results within the

reporting deadline Three non-NRLs did not submit any results two of which (L21 L28)

had not received the samples from customs while L67 did not provide any justification for

not participating

Table 3 shows an overview of the participation to this PT

Table 3 Communication about and participation to the PT 0217

Characteristic of the PT Result

Date of PT announcement 31 May 2017

Deadline for registration 14 June 2017

Date of shipment of samples 3 July 2017

Deadline for result submission 25 August 2017

Number of laboratories informed 199

Number of registered laboratories 89

Registered laboratories that failed to submit their data 3

Number of participating laboratories 86

The participating laboratories fell into the following assigned categories (Table 4)

Thirty-three NRLs designated in line with Regulation (EC) No 8822004 (NRL882)

representing 25 EU Member States (many of them are also NRL120) In addition

Ireland delegated its NRL882 tasks to one of the PT participants Estonia and

Malta were not represented in this PT

Twenty-one NRLs nominated under Regulation (EU) No 1202014 (NRL120) that

are not at the same time NRL under Regulation (EC) No 8822004

Thirty-two laboratories that are not NRL but are appointed by their National

Authority to perform official controls This category includes 10 EU official control

laboratories (OCLs) and 22 laboratories from non-EU countries including Serbia

and Switzerland

Among the countries Germany was represented with 17 laboratories Italy with 6

laboratories and Belgium and Poland with 4 laboratories each all other countries had

between one and three participating laboratories

EURL GMFF CT 0217 report

1346

Table 4 Overview of participants to CT 0217 by country and category

Country Participants NRL8821 NRL120 Non-NRL

AUSTRIA 2 2

BELGIUM 4 3

1

BRAZIL 2

2

BULGARIA 2 1

1

CHILE 1

1

COLOMBIA 1

1

CROATIA 2 1

1

CYPRUS 1 1

CZECH REPUBLIC 1 1

DENMARK 1 1

FINLAND 2 1 1

FRANCE 2 2

GERMANY 17 1 14 2

GREECE 1 1

HUNGARY 2 1

1

INDIA 1

1

ITALY 6 1 2 3

LATVIA 1 1

LEBANON 1

1

LITHUANIA 1 1

LUXEMBOURG 1 1

MEXICO 1

1

NETHERLANDS 2 1 1

PHILIPPINES 1

1

POLAND 4 3 1

PORTUGAL 1 1

ROMANIA 2 1

1

SERBIA 3

3

SINGAPORE 1

1

SLOVAKIA 2 2

SLOVENIA 1 1

SPAIN 2 2

SWEDEN 1 1

SWITZERLAND 2

2

TURKEY 1

1

UKRAINE 3

3

UNITED KINGDOM 3 1 2

UNITED STATES 1

1

VIETNAM 3

3

Total 86 33 21 32 1 No NRL882 from Estonia or Malta participated to this PT

EURL GMFF CT 0217 report

1446

42 Information on the testing provided in the questionnaire

Participants were asked to fill in an online questionnaire (through EUSurvey) on their

testing methodology used for T1 and T2 consisting of a number of mostly multiple-choice

questions A total of 82 laboratories completed the questionnaire including all 33

NRL882 20 out of 21 NRL120 and 29 out of 32 non-NRLs (questionnaires were missing

from L03 L10 L38 and L49)

Table 5 summarises the main answers received whereas Annex 2 shows all answers The

results on GM event identification are reported in Section 43

Table 5 Summary of the main answers provided in the questionnaire of CT 0217

Question (and Question

number) Test Item 1 ndash 40-3-2 Test Item 2 ndash 40-3-2

Test item analysed Yes (801) No (2) Yes (82) No (0)

Reason for lack of analysis (Q1)

Matrix out of scope (1) other practical constraints (1) -

DNA extraction method (Q2) CTAB (41) NucleoSpin Food (10) CTAB (40) NucleoSpin Food (11)

Additional DNA purification method (Q3)

None (51) Ethanol (9) None (52) Ethanol (9)

Number of replicates (Q4) 2 (52) 4 (10) 2 (56) 4 (10)

Approach to test for PCR inhibition (Q5)

OD ratios (40) delta Cq or GM between two dilutions (31)

OD ratios (36) delta Cq or GM between two dilutions (32)

Reason for not testing all events (Q9)

Not applicable (45) below the LOQ (18) Not applicable (61) reagents not available (10)

Approach used (Q6a) Standard curves (68) delta Cq (9) Standard curves (71) delta Cq (9)

Calibrant used (Q6b) CRM JRC-Geel in gkg (71) other RM in copies (4) CRM JRC-Geel in gkg (74) other RM in copies (4)

Taxon-specific endogenous gene (Q6c)

lec-74 bp (65) lec-81 or 118 bp (each 5) lec-74 bp (63) lec-118 bp (6)

Unit of measurement and data expression (Q6d)

Mass (59) copies=mass CRM (13) Mass (61) copies=mass CRM (14)

Amount of DNA (Q6e) 200 ng (30) 100 ng (18) 200 ng (30) 100 ng (21)

LOQ (Q6f) 01 (37) lt01 (34) 01 (39) lt01 (39)

LOQ determination (Q6g) In-house validation (42) current analysis (20) In-house validation (43) current analysis (21)

Uncertainty determination (Q6h)

Precision of replicates (32) in-house validation (27) Precision of replicates (32) in-house validation (30)

1 The numbers shown refer to the number of laboratories that reported the answer The answers that were reported with the two largest frequencies are mentioned

One NRL882 (L73) reported that T1 was out of the scope of the laboratory and one non-

NRL (L43) reported that the T1 matrix was not analysed because of practical constraints

The evaluation of the answers shows that the most commonly employed DNA extraction

method for both T1 and T2 was one based on CTAB with the NucleoSpin Food kit ranking

second No additional purification methods were generally applied The majority of

laboratories analysed two replicate DNA extracts Most laboratories checked the quality of

the DNA extracts by verifying the OD ratios andor running two dilutions a minority of

laboratories performed a PCR inhibition run on 3 or 4 DNA dilutions with a reference gene

For the quantitative analysis the most common approach used was based on two standard

curves however 9 laboratories applied the delta Cq approach One laboratory (L50)

mentioned the use of digital PCR for 40-3-2 soybean quantification in T1 and T2 The

available CRMs from the JRC were used by most laboratories but 5 laboratories used a

non-certified reference material (RM) where values were expressed in GM copy number

ratio (4) or GM mass fraction (1) Lec was used as taxon-specific reference gene by all

laboratories for soybean (mostly the 74 bp version) The majority of laboratories

performed their measurements in the same unit as the certified value of the calibrant used

(gkg) and no conversion factor was applied The LOQ reported was either taken from in-

house validation of the method or determined from the analysis results for this CT In

most cases a LOQ of 01 mm or lower was reported The measurement uncertainty

was either derived from the standard deviation of the measurement replicates or from the

intermediate precision determined in the frame of the single-laboratory validation study

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

1146

3 Instructions to the participants

Participants in this PT were instructed to analyse the two test items (T1 and T2) as

follows

Test Item 1 Chicken feed

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Test Item 2 Soybean flour

- Screen for the presence of the following three GM soybean event(s) 40-3-2

68416 and MON89788

- Quantify the GM soybean event that is present with the highest GM mass fraction

Quantitative results had to be reported in mm as outlined below

Mass GM event [g]

mm = x 100 (1)

Total mass species [g]

Participants were requested

- to use procedures for detectionquantification of the GM events that resemble as

closely as possible the ones used in routine sample analysis

- to take care in ensuring that the DNA extraction method employed is adapted to

the matrix and that the quality of the DNA obtained is suitable for PCR

- to report the quantitative results with two decimal places (eg 064 or 129)

- to follow the general rule that results obtained using a calibrant certified for GM

mass fraction (ie a matrix CRM certified in [x] gkg) can directly be expressed in

mm while results obtained using a calibrant certified for DNA copy number

ratio (eg a plasmid containing both the GM and reference gene target or some

matrix CRMs) need to be converted into mm using a conversion factor(1011)

- to pay attention to the correct estimation and reporting of the measurement

uncertainty and coverage factor used as the uncertainty reported would be

considered in the evaluation of the results using zeta scores

- in case of an unsatisfactory performance to fill in a form indicating the root-cause

analysis and providing evidence demonstrating the effectiveness of the corrective

actions implemented (for NRLs only)

The participants were also informed that the identification information on the participants

in this PT would be kept confidential except for the NRLs that have been appointed in line

with Regulation (EC) No 8822004 their lab codes will be disclosed to DG SANTE for the

purpose of an assessment of their performance

EURL GMFF CT 0217 report

1246

4 Results

41 Participation to the PT

On 31 May 2017 199 laboratories were informed about the upcoming PT EURL-GMFF-CT-

0217 Finally 89 laboratories registered for it and received a random unique lab code

(L01 to L89) Eighty-six laboratories from 39 countries returned results within the

reporting deadline Three non-NRLs did not submit any results two of which (L21 L28)

had not received the samples from customs while L67 did not provide any justification for

not participating

Table 3 shows an overview of the participation to this PT

Table 3 Communication about and participation to the PT 0217

Characteristic of the PT Result

Date of PT announcement 31 May 2017

Deadline for registration 14 June 2017

Date of shipment of samples 3 July 2017

Deadline for result submission 25 August 2017

Number of laboratories informed 199

Number of registered laboratories 89

Registered laboratories that failed to submit their data 3

Number of participating laboratories 86

The participating laboratories fell into the following assigned categories (Table 4)

Thirty-three NRLs designated in line with Regulation (EC) No 8822004 (NRL882)

representing 25 EU Member States (many of them are also NRL120) In addition

Ireland delegated its NRL882 tasks to one of the PT participants Estonia and

Malta were not represented in this PT

Twenty-one NRLs nominated under Regulation (EU) No 1202014 (NRL120) that

are not at the same time NRL under Regulation (EC) No 8822004

Thirty-two laboratories that are not NRL but are appointed by their National

Authority to perform official controls This category includes 10 EU official control

laboratories (OCLs) and 22 laboratories from non-EU countries including Serbia

and Switzerland

Among the countries Germany was represented with 17 laboratories Italy with 6

laboratories and Belgium and Poland with 4 laboratories each all other countries had

between one and three participating laboratories

EURL GMFF CT 0217 report

1346

Table 4 Overview of participants to CT 0217 by country and category

Country Participants NRL8821 NRL120 Non-NRL

AUSTRIA 2 2

BELGIUM 4 3

1

BRAZIL 2

2

BULGARIA 2 1

1

CHILE 1

1

COLOMBIA 1

1

CROATIA 2 1

1

CYPRUS 1 1

CZECH REPUBLIC 1 1

DENMARK 1 1

FINLAND 2 1 1

FRANCE 2 2

GERMANY 17 1 14 2

GREECE 1 1

HUNGARY 2 1

1

INDIA 1

1

ITALY 6 1 2 3

LATVIA 1 1

LEBANON 1

1

LITHUANIA 1 1

LUXEMBOURG 1 1

MEXICO 1

1

NETHERLANDS 2 1 1

PHILIPPINES 1

1

POLAND 4 3 1

PORTUGAL 1 1

ROMANIA 2 1

1

SERBIA 3

3

SINGAPORE 1

1

SLOVAKIA 2 2

SLOVENIA 1 1

SPAIN 2 2

SWEDEN 1 1

SWITZERLAND 2

2

TURKEY 1

1

UKRAINE 3

3

UNITED KINGDOM 3 1 2

UNITED STATES 1

1

VIETNAM 3

3

Total 86 33 21 32 1 No NRL882 from Estonia or Malta participated to this PT

EURL GMFF CT 0217 report

1446

42 Information on the testing provided in the questionnaire

Participants were asked to fill in an online questionnaire (through EUSurvey) on their

testing methodology used for T1 and T2 consisting of a number of mostly multiple-choice

questions A total of 82 laboratories completed the questionnaire including all 33

NRL882 20 out of 21 NRL120 and 29 out of 32 non-NRLs (questionnaires were missing

from L03 L10 L38 and L49)

Table 5 summarises the main answers received whereas Annex 2 shows all answers The

results on GM event identification are reported in Section 43

Table 5 Summary of the main answers provided in the questionnaire of CT 0217

Question (and Question

number) Test Item 1 ndash 40-3-2 Test Item 2 ndash 40-3-2

Test item analysed Yes (801) No (2) Yes (82) No (0)

Reason for lack of analysis (Q1)

Matrix out of scope (1) other practical constraints (1) -

DNA extraction method (Q2) CTAB (41) NucleoSpin Food (10) CTAB (40) NucleoSpin Food (11)

Additional DNA purification method (Q3)

None (51) Ethanol (9) None (52) Ethanol (9)

Number of replicates (Q4) 2 (52) 4 (10) 2 (56) 4 (10)

Approach to test for PCR inhibition (Q5)

OD ratios (40) delta Cq or GM between two dilutions (31)

OD ratios (36) delta Cq or GM between two dilutions (32)

Reason for not testing all events (Q9)

Not applicable (45) below the LOQ (18) Not applicable (61) reagents not available (10)

Approach used (Q6a) Standard curves (68) delta Cq (9) Standard curves (71) delta Cq (9)

Calibrant used (Q6b) CRM JRC-Geel in gkg (71) other RM in copies (4) CRM JRC-Geel in gkg (74) other RM in copies (4)

Taxon-specific endogenous gene (Q6c)

lec-74 bp (65) lec-81 or 118 bp (each 5) lec-74 bp (63) lec-118 bp (6)

Unit of measurement and data expression (Q6d)

Mass (59) copies=mass CRM (13) Mass (61) copies=mass CRM (14)

Amount of DNA (Q6e) 200 ng (30) 100 ng (18) 200 ng (30) 100 ng (21)

LOQ (Q6f) 01 (37) lt01 (34) 01 (39) lt01 (39)

LOQ determination (Q6g) In-house validation (42) current analysis (20) In-house validation (43) current analysis (21)

Uncertainty determination (Q6h)

Precision of replicates (32) in-house validation (27) Precision of replicates (32) in-house validation (30)

1 The numbers shown refer to the number of laboratories that reported the answer The answers that were reported with the two largest frequencies are mentioned

One NRL882 (L73) reported that T1 was out of the scope of the laboratory and one non-

NRL (L43) reported that the T1 matrix was not analysed because of practical constraints

The evaluation of the answers shows that the most commonly employed DNA extraction

method for both T1 and T2 was one based on CTAB with the NucleoSpin Food kit ranking

second No additional purification methods were generally applied The majority of

laboratories analysed two replicate DNA extracts Most laboratories checked the quality of

the DNA extracts by verifying the OD ratios andor running two dilutions a minority of

laboratories performed a PCR inhibition run on 3 or 4 DNA dilutions with a reference gene

For the quantitative analysis the most common approach used was based on two standard

curves however 9 laboratories applied the delta Cq approach One laboratory (L50)

mentioned the use of digital PCR for 40-3-2 soybean quantification in T1 and T2 The

available CRMs from the JRC were used by most laboratories but 5 laboratories used a

non-certified reference material (RM) where values were expressed in GM copy number

ratio (4) or GM mass fraction (1) Lec was used as taxon-specific reference gene by all

laboratories for soybean (mostly the 74 bp version) The majority of laboratories

performed their measurements in the same unit as the certified value of the calibrant used

(gkg) and no conversion factor was applied The LOQ reported was either taken from in-

house validation of the method or determined from the analysis results for this CT In

most cases a LOQ of 01 mm or lower was reported The measurement uncertainty

was either derived from the standard deviation of the measurement replicates or from the

intermediate precision determined in the frame of the single-laboratory validation study

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

1246

4 Results

41 Participation to the PT

On 31 May 2017 199 laboratories were informed about the upcoming PT EURL-GMFF-CT-

0217 Finally 89 laboratories registered for it and received a random unique lab code

(L01 to L89) Eighty-six laboratories from 39 countries returned results within the

reporting deadline Three non-NRLs did not submit any results two of which (L21 L28)

had not received the samples from customs while L67 did not provide any justification for

not participating

Table 3 shows an overview of the participation to this PT

Table 3 Communication about and participation to the PT 0217

Characteristic of the PT Result

Date of PT announcement 31 May 2017

Deadline for registration 14 June 2017

Date of shipment of samples 3 July 2017

Deadline for result submission 25 August 2017

Number of laboratories informed 199

Number of registered laboratories 89

Registered laboratories that failed to submit their data 3

Number of participating laboratories 86

The participating laboratories fell into the following assigned categories (Table 4)

Thirty-three NRLs designated in line with Regulation (EC) No 8822004 (NRL882)

representing 25 EU Member States (many of them are also NRL120) In addition

Ireland delegated its NRL882 tasks to one of the PT participants Estonia and

Malta were not represented in this PT

Twenty-one NRLs nominated under Regulation (EU) No 1202014 (NRL120) that

are not at the same time NRL under Regulation (EC) No 8822004

Thirty-two laboratories that are not NRL but are appointed by their National

Authority to perform official controls This category includes 10 EU official control

laboratories (OCLs) and 22 laboratories from non-EU countries including Serbia

and Switzerland

Among the countries Germany was represented with 17 laboratories Italy with 6

laboratories and Belgium and Poland with 4 laboratories each all other countries had

between one and three participating laboratories

EURL GMFF CT 0217 report

1346

Table 4 Overview of participants to CT 0217 by country and category

Country Participants NRL8821 NRL120 Non-NRL

AUSTRIA 2 2

BELGIUM 4 3

1

BRAZIL 2

2

BULGARIA 2 1

1

CHILE 1

1

COLOMBIA 1

1

CROATIA 2 1

1

CYPRUS 1 1

CZECH REPUBLIC 1 1

DENMARK 1 1

FINLAND 2 1 1

FRANCE 2 2

GERMANY 17 1 14 2

GREECE 1 1

HUNGARY 2 1

1

INDIA 1

1

ITALY 6 1 2 3

LATVIA 1 1

LEBANON 1

1

LITHUANIA 1 1

LUXEMBOURG 1 1

MEXICO 1

1

NETHERLANDS 2 1 1

PHILIPPINES 1

1

POLAND 4 3 1

PORTUGAL 1 1

ROMANIA 2 1

1

SERBIA 3

3

SINGAPORE 1

1

SLOVAKIA 2 2

SLOVENIA 1 1

SPAIN 2 2

SWEDEN 1 1

SWITZERLAND 2

2

TURKEY 1

1

UKRAINE 3

3

UNITED KINGDOM 3 1 2

UNITED STATES 1

1

VIETNAM 3

3

Total 86 33 21 32 1 No NRL882 from Estonia or Malta participated to this PT

EURL GMFF CT 0217 report

1446

42 Information on the testing provided in the questionnaire

Participants were asked to fill in an online questionnaire (through EUSurvey) on their

testing methodology used for T1 and T2 consisting of a number of mostly multiple-choice

questions A total of 82 laboratories completed the questionnaire including all 33

NRL882 20 out of 21 NRL120 and 29 out of 32 non-NRLs (questionnaires were missing

from L03 L10 L38 and L49)

Table 5 summarises the main answers received whereas Annex 2 shows all answers The

results on GM event identification are reported in Section 43

Table 5 Summary of the main answers provided in the questionnaire of CT 0217

Question (and Question

number) Test Item 1 ndash 40-3-2 Test Item 2 ndash 40-3-2

Test item analysed Yes (801) No (2) Yes (82) No (0)

Reason for lack of analysis (Q1)

Matrix out of scope (1) other practical constraints (1) -

DNA extraction method (Q2) CTAB (41) NucleoSpin Food (10) CTAB (40) NucleoSpin Food (11)

Additional DNA purification method (Q3)

None (51) Ethanol (9) None (52) Ethanol (9)

Number of replicates (Q4) 2 (52) 4 (10) 2 (56) 4 (10)

Approach to test for PCR inhibition (Q5)

OD ratios (40) delta Cq or GM between two dilutions (31)

OD ratios (36) delta Cq or GM between two dilutions (32)

Reason for not testing all events (Q9)

Not applicable (45) below the LOQ (18) Not applicable (61) reagents not available (10)

Approach used (Q6a) Standard curves (68) delta Cq (9) Standard curves (71) delta Cq (9)

Calibrant used (Q6b) CRM JRC-Geel in gkg (71) other RM in copies (4) CRM JRC-Geel in gkg (74) other RM in copies (4)

Taxon-specific endogenous gene (Q6c)

lec-74 bp (65) lec-81 or 118 bp (each 5) lec-74 bp (63) lec-118 bp (6)

Unit of measurement and data expression (Q6d)

Mass (59) copies=mass CRM (13) Mass (61) copies=mass CRM (14)

Amount of DNA (Q6e) 200 ng (30) 100 ng (18) 200 ng (30) 100 ng (21)

LOQ (Q6f) 01 (37) lt01 (34) 01 (39) lt01 (39)

LOQ determination (Q6g) In-house validation (42) current analysis (20) In-house validation (43) current analysis (21)

Uncertainty determination (Q6h)

Precision of replicates (32) in-house validation (27) Precision of replicates (32) in-house validation (30)

1 The numbers shown refer to the number of laboratories that reported the answer The answers that were reported with the two largest frequencies are mentioned

One NRL882 (L73) reported that T1 was out of the scope of the laboratory and one non-

NRL (L43) reported that the T1 matrix was not analysed because of practical constraints

The evaluation of the answers shows that the most commonly employed DNA extraction

method for both T1 and T2 was one based on CTAB with the NucleoSpin Food kit ranking

second No additional purification methods were generally applied The majority of

laboratories analysed two replicate DNA extracts Most laboratories checked the quality of

the DNA extracts by verifying the OD ratios andor running two dilutions a minority of

laboratories performed a PCR inhibition run on 3 or 4 DNA dilutions with a reference gene

For the quantitative analysis the most common approach used was based on two standard

curves however 9 laboratories applied the delta Cq approach One laboratory (L50)

mentioned the use of digital PCR for 40-3-2 soybean quantification in T1 and T2 The

available CRMs from the JRC were used by most laboratories but 5 laboratories used a

non-certified reference material (RM) where values were expressed in GM copy number

ratio (4) or GM mass fraction (1) Lec was used as taxon-specific reference gene by all

laboratories for soybean (mostly the 74 bp version) The majority of laboratories

performed their measurements in the same unit as the certified value of the calibrant used

(gkg) and no conversion factor was applied The LOQ reported was either taken from in-

house validation of the method or determined from the analysis results for this CT In

most cases a LOQ of 01 mm or lower was reported The measurement uncertainty

was either derived from the standard deviation of the measurement replicates or from the

intermediate precision determined in the frame of the single-laboratory validation study

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

1346

Table 4 Overview of participants to CT 0217 by country and category

Country Participants NRL8821 NRL120 Non-NRL

AUSTRIA 2 2

BELGIUM 4 3

1

BRAZIL 2

2

BULGARIA 2 1

1

CHILE 1

1

COLOMBIA 1

1

CROATIA 2 1

1

CYPRUS 1 1

CZECH REPUBLIC 1 1

DENMARK 1 1

FINLAND 2 1 1

FRANCE 2 2

GERMANY 17 1 14 2

GREECE 1 1

HUNGARY 2 1

1

INDIA 1

1

ITALY 6 1 2 3

LATVIA 1 1

LEBANON 1

1

LITHUANIA 1 1

LUXEMBOURG 1 1

MEXICO 1

1

NETHERLANDS 2 1 1

PHILIPPINES 1

1

POLAND 4 3 1

PORTUGAL 1 1

ROMANIA 2 1

1

SERBIA 3

3

SINGAPORE 1

1

SLOVAKIA 2 2

SLOVENIA 1 1

SPAIN 2 2

SWEDEN 1 1

SWITZERLAND 2

2

TURKEY 1

1

UKRAINE 3

3

UNITED KINGDOM 3 1 2

UNITED STATES 1

1

VIETNAM 3

3

Total 86 33 21 32 1 No NRL882 from Estonia or Malta participated to this PT

EURL GMFF CT 0217 report

1446

42 Information on the testing provided in the questionnaire

Participants were asked to fill in an online questionnaire (through EUSurvey) on their

testing methodology used for T1 and T2 consisting of a number of mostly multiple-choice

questions A total of 82 laboratories completed the questionnaire including all 33

NRL882 20 out of 21 NRL120 and 29 out of 32 non-NRLs (questionnaires were missing

from L03 L10 L38 and L49)

Table 5 summarises the main answers received whereas Annex 2 shows all answers The

results on GM event identification are reported in Section 43

Table 5 Summary of the main answers provided in the questionnaire of CT 0217

Question (and Question

number) Test Item 1 ndash 40-3-2 Test Item 2 ndash 40-3-2

Test item analysed Yes (801) No (2) Yes (82) No (0)

Reason for lack of analysis (Q1)

Matrix out of scope (1) other practical constraints (1) -

DNA extraction method (Q2) CTAB (41) NucleoSpin Food (10) CTAB (40) NucleoSpin Food (11)

Additional DNA purification method (Q3)

None (51) Ethanol (9) None (52) Ethanol (9)

Number of replicates (Q4) 2 (52) 4 (10) 2 (56) 4 (10)

Approach to test for PCR inhibition (Q5)

OD ratios (40) delta Cq or GM between two dilutions (31)

OD ratios (36) delta Cq or GM between two dilutions (32)

Reason for not testing all events (Q9)

Not applicable (45) below the LOQ (18) Not applicable (61) reagents not available (10)

Approach used (Q6a) Standard curves (68) delta Cq (9) Standard curves (71) delta Cq (9)

Calibrant used (Q6b) CRM JRC-Geel in gkg (71) other RM in copies (4) CRM JRC-Geel in gkg (74) other RM in copies (4)

Taxon-specific endogenous gene (Q6c)

lec-74 bp (65) lec-81 or 118 bp (each 5) lec-74 bp (63) lec-118 bp (6)

Unit of measurement and data expression (Q6d)

Mass (59) copies=mass CRM (13) Mass (61) copies=mass CRM (14)

Amount of DNA (Q6e) 200 ng (30) 100 ng (18) 200 ng (30) 100 ng (21)

LOQ (Q6f) 01 (37) lt01 (34) 01 (39) lt01 (39)

LOQ determination (Q6g) In-house validation (42) current analysis (20) In-house validation (43) current analysis (21)

Uncertainty determination (Q6h)

Precision of replicates (32) in-house validation (27) Precision of replicates (32) in-house validation (30)

1 The numbers shown refer to the number of laboratories that reported the answer The answers that were reported with the two largest frequencies are mentioned

One NRL882 (L73) reported that T1 was out of the scope of the laboratory and one non-

NRL (L43) reported that the T1 matrix was not analysed because of practical constraints

The evaluation of the answers shows that the most commonly employed DNA extraction

method for both T1 and T2 was one based on CTAB with the NucleoSpin Food kit ranking

second No additional purification methods were generally applied The majority of

laboratories analysed two replicate DNA extracts Most laboratories checked the quality of

the DNA extracts by verifying the OD ratios andor running two dilutions a minority of

laboratories performed a PCR inhibition run on 3 or 4 DNA dilutions with a reference gene

For the quantitative analysis the most common approach used was based on two standard

curves however 9 laboratories applied the delta Cq approach One laboratory (L50)

mentioned the use of digital PCR for 40-3-2 soybean quantification in T1 and T2 The

available CRMs from the JRC were used by most laboratories but 5 laboratories used a

non-certified reference material (RM) where values were expressed in GM copy number

ratio (4) or GM mass fraction (1) Lec was used as taxon-specific reference gene by all

laboratories for soybean (mostly the 74 bp version) The majority of laboratories

performed their measurements in the same unit as the certified value of the calibrant used

(gkg) and no conversion factor was applied The LOQ reported was either taken from in-

house validation of the method or determined from the analysis results for this CT In

most cases a LOQ of 01 mm or lower was reported The measurement uncertainty

was either derived from the standard deviation of the measurement replicates or from the

intermediate precision determined in the frame of the single-laboratory validation study

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

1446

42 Information on the testing provided in the questionnaire

Participants were asked to fill in an online questionnaire (through EUSurvey) on their

testing methodology used for T1 and T2 consisting of a number of mostly multiple-choice

questions A total of 82 laboratories completed the questionnaire including all 33

NRL882 20 out of 21 NRL120 and 29 out of 32 non-NRLs (questionnaires were missing

from L03 L10 L38 and L49)

Table 5 summarises the main answers received whereas Annex 2 shows all answers The

results on GM event identification are reported in Section 43

Table 5 Summary of the main answers provided in the questionnaire of CT 0217

Question (and Question

number) Test Item 1 ndash 40-3-2 Test Item 2 ndash 40-3-2

Test item analysed Yes (801) No (2) Yes (82) No (0)

Reason for lack of analysis (Q1)

Matrix out of scope (1) other practical constraints (1) -

DNA extraction method (Q2) CTAB (41) NucleoSpin Food (10) CTAB (40) NucleoSpin Food (11)

Additional DNA purification method (Q3)

None (51) Ethanol (9) None (52) Ethanol (9)

Number of replicates (Q4) 2 (52) 4 (10) 2 (56) 4 (10)

Approach to test for PCR inhibition (Q5)

OD ratios (40) delta Cq or GM between two dilutions (31)

OD ratios (36) delta Cq or GM between two dilutions (32)

Reason for not testing all events (Q9)

Not applicable (45) below the LOQ (18) Not applicable (61) reagents not available (10)

Approach used (Q6a) Standard curves (68) delta Cq (9) Standard curves (71) delta Cq (9)

Calibrant used (Q6b) CRM JRC-Geel in gkg (71) other RM in copies (4) CRM JRC-Geel in gkg (74) other RM in copies (4)

Taxon-specific endogenous gene (Q6c)

lec-74 bp (65) lec-81 or 118 bp (each 5) lec-74 bp (63) lec-118 bp (6)

Unit of measurement and data expression (Q6d)

Mass (59) copies=mass CRM (13) Mass (61) copies=mass CRM (14)

Amount of DNA (Q6e) 200 ng (30) 100 ng (18) 200 ng (30) 100 ng (21)

LOQ (Q6f) 01 (37) lt01 (34) 01 (39) lt01 (39)

LOQ determination (Q6g) In-house validation (42) current analysis (20) In-house validation (43) current analysis (21)

Uncertainty determination (Q6h)

Precision of replicates (32) in-house validation (27) Precision of replicates (32) in-house validation (30)

1 The numbers shown refer to the number of laboratories that reported the answer The answers that were reported with the two largest frequencies are mentioned

One NRL882 (L73) reported that T1 was out of the scope of the laboratory and one non-

NRL (L43) reported that the T1 matrix was not analysed because of practical constraints

The evaluation of the answers shows that the most commonly employed DNA extraction

method for both T1 and T2 was one based on CTAB with the NucleoSpin Food kit ranking

second No additional purification methods were generally applied The majority of

laboratories analysed two replicate DNA extracts Most laboratories checked the quality of

the DNA extracts by verifying the OD ratios andor running two dilutions a minority of

laboratories performed a PCR inhibition run on 3 or 4 DNA dilutions with a reference gene

For the quantitative analysis the most common approach used was based on two standard

curves however 9 laboratories applied the delta Cq approach One laboratory (L50)

mentioned the use of digital PCR for 40-3-2 soybean quantification in T1 and T2 The

available CRMs from the JRC were used by most laboratories but 5 laboratories used a

non-certified reference material (RM) where values were expressed in GM copy number

ratio (4) or GM mass fraction (1) Lec was used as taxon-specific reference gene by all

laboratories for soybean (mostly the 74 bp version) The majority of laboratories

performed their measurements in the same unit as the certified value of the calibrant used

(gkg) and no conversion factor was applied The LOQ reported was either taken from in-

house validation of the method or determined from the analysis results for this CT In

most cases a LOQ of 01 mm or lower was reported The measurement uncertainty

was either derived from the standard deviation of the measurement replicates or from the

intermediate precision determined in the frame of the single-laboratory validation study

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

1546

43 GM event identification

Table 6 summarises the results reported by the participants through the questionnaire

regarding the (qualitative) identification of the GM events

Table 6 Summary of GM event identification results of the participants as reported in the

questionnaire or (in brackets) inferred from the quantitative result reported

Laboratories Test Item GM Event Present Absent Not Tested Sample Not Analysed

NRL882 and NRL120

T1

40-3-2 52 (+1) 0 0

1 68416 0 51 1

MON89788 32 19 1

T2

40-3-2 53 (+1) 0 0

0 68416 0 52 1

MON89788 5 47 1

Non-NRLs

T1

40-3-2 28 (+2) 0 0

1 68416 1 14 13

MON89788 16 7 5

T2

40-3-2 29 (+2) 0 0

0 68416 1 14 14

MON89788 4 21 4

All 53 NRLs who had tested T1 identified the 40-3-2 event in T1 The 68416 soybean event

was found absent in T1 whereas MON89788 soybean was detected by 32 NRLs in T1

Seven NRLs also reported to have quantified MON89788 in T1 and while most laboratories

reported the GM mass fraction as being below the LOQ two laboratories reported a GM

mass fraction of 002 and 0025 mm The presence of traces of MON89788 soybean in

the chicken feed was indeed confirmed by JRC-Geel (see Table 1)

For T2 all 54 NRLs identified the 40-3-2 event and found 68416 absent MON89788

soybean was detected by 5 NRLs but quantified as below the LOQ (note that the presence

of MON89788 in T2 was not confirmed by JRC-Geel) The results show that EU NRLs are

able to correctly identify the 40-3-2 soybean event in both a compound feed matrix and in

soybean flour

The results of all non-NRLs were also satisfactory for event 40-3-2 soybean however a

larger proportion of laboratories did not test the event MON89788 and particularly

68416

The performance of all laboratories for qualitative identification of the correct GM events is

summarised in Annex 3

44 GM event quantification

441 Number of participants reporting a quantitative result

Table 7 presents the number of laboratories having submitted quantitative data for the GM

event present in the test items A large majority of participating laboratories reported a

quantitative result for 40-3-2 soybean in T1 (93 ) and T2 (97 ) Among the NRLs one

NRL120 (L88) provided a result for 40-3-2 soybean in T2 but not for T1 reporting that

the mass fraction of the 40-3-2 event was below the LOQ in T1 All NRL882 participants

quantified the event in both test items except L73 for which the T1 matrix was out of

their scope

Expanded measurement uncertainties were reported by the NRLs for all measurement

results with the coverage factor reported for 88 and 87 of the results for T1 and T2

respectively (Table 7) Although the results show that most control laboratories

understand the principle that analytical results should be reported with an expanded

uncertainty when asked it is unclear why some laboratories did not report the coverage

factor (k) used to convert the standard uncertainty into an expanded uncertainty that

corresponds to a 95 level of confidence

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

1646

Table 7 Number of laboratories reporting a quantitative GM event-specific result

Quantitative Results

Reported

Test Item 1 ndash 40-3-2 Soybean Test Item 2 ndash 40-3-2 Soybean

NRL882 NRL120 Non-NRL NRL882 NRL120 Non-NRL

Total participants 33 21 32 33 21 32

Quantitative result 32a 20b 28c 33 21 29c

Measurement uncertainty 32 20 19 33 21 20

Coverage factor 28 18 17 28 19 18 a L73 (NRL882) did not analyse T1 as the matrix is out of the scope of the laboratory b L88 (NRL120) did not provide a quantitative result for 40-3-2 soybean in T1 c L10 L11 and L71 (non-NRLs) have not reported a quantitative result for T1 and T2 L43 has not analysed T1

442 Assigned values

The assigned value (xpt) for the mass fraction of event 40-3-2 soybean in T1 and T2 was

based on the consensus value of the data from a pre-selected part of participants in this

PT calculated using robust statistics(1213) This statistical approach minimises the

influence of outlying values The data taken into account for the calculation of the robust

means were those from the NRLs (NRL882 and NRL120) only The data from non-NRLs

were excluded because of the heterogeneity of this group with regard to experience in

GMO analysis

The results of proficiency tests for the analysis of GMOs are often log-normally distributed

(skewed)(1415) This was not the case for the results of this PT however for consistency

with previous PTs the same approach was followed as in previous rounds The results

reported by the NRLs were first log10-transformed and the robust mean (xpt-log) and

corresponding robust standard deviation (s) were calculated The standard measurement

uncertainty [u(xpt-log)] of the assigned value is assumed to include the effects of

uncertainty due to inhomogeneity and instability it is estimated according to ISO

135282015 (section 773)(16) as follows

N

s )x(u

logpt 251 (2)

where s = robust standard deviation of the results expressed in mm (log scale)

N = number of results used for the calculation (from NRLs only)

A coverage factor (k) of 2 was used to calculate the expanded uncertainty (U)

corresponding to a 95 level of confidence(17)

The assigned values and associated uncertainties for 40-3-2 soybean in both test items

are reported in Table 8 The standard deviation for proficiency assessment (σpt-log) was set

to 010 (on the log scale) for both test items based on reasonable performance

expectations and experience from previous PTs

Table 8 Overview of assigned values and uncertainties for the GM mass fraction in T1 and T2

Variable 40-3-2 Soybean in T1 40-3-2 Soybean in T2

Assigned value derived as Robust mean of log10-transformed data Robust mean of log10-transformed data

Number of data points (NRLs) 52 54

Assigned Value (xpt-log) -00973a -01186b

Standard uncertainty [u(xpt-log)] 00207 00122

Standard deviation for proficiency assessment (σpt-log)

010 010

a The assigned value for the mass fraction of 40-3-2 soybean in T1 corresponds to an approximate GM in the raw domain of 080 mm b The assigned value for the mass fraction of 40-3-2 soybean in T2 corresponds to an approximate GM in the raw domain of 076 mm

The robust mean of the 40-3-2 soybean mass fraction reported for T2 was 24 lower (on

the raw scale) than the certified value of this new batch of CRM(6) which was released in

October 2017 (ie after this PT) This observation was already noticed during CRM

production and a note on this is included in the CRM certificate (see Section 22 in this PT

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

1746

report) Laboratories that have calibrated their measurements with the previous 40-3-2

soybean CRM (ERM-BF410n series) or another RM characterised by a 11 ratio between

the DNA content of the GM and non-GM fraction are therefore expected to measure a

much lower GM mass fraction in the new CRM than the certified value of 100 mm The

important message here is to not mix up the previous CRM batch with the new one as

both batches do not contain the same amount of transgenic copies Despite this

observation the aim of using a CRM calibrant is to have a common reference point for the

implementation of EU legislation on GMO thresholds and labelling(11)

443 Calculation of performance scores

Individual laboratory performance was expressed in terms of z and ζ scores in accordance

with ISO 135282015(16) both calculated in the log domain as follows

logpt

logpti

σ

xx

)log(z (4)

22 )x(u)x(u

)log(

logptlogi

logpti xx (5)

where xi = the measurement result as reported by a participant

u(xi) = the standard measurement uncertainty of the result reported xpt-log = the assigned value

u(xi-log) = the standard measurement uncertainty of the result reported

u(xpt-log) = the standard measurement uncertainty of the assigned value

σpt-log = the standard deviation for proficiency assessment

For calculation of the ζ scores the expanded uncertainties on the results reported by the

laboratories were translated into standard measurement uncertainties [u(xi)] using the

coverage factor reported and converted to the log domain as follows (following general

rules for the measurement uncertainty of log10-transformed values)

i

i

x

x )(u )x(u logi 4340 (6)

When no measurement uncertainty was reported it was set to zero (u(xi) = 0) When no

coverage factor was reported k was set to 173 (assuming a rectangular distribution

around the reported value with boundaries valuing plusmn Uradic3 (17))

Performance scores were calculated on the results as reported by the participants and

rounded to one decimal afterwards The interpretation of the z and ζ performance scores

was done according to ISO 170432010(5)

|score| le 20 satisfactory performance

20 lt |score| lt 30 questionable performance

|score| ge 30 unsatisfactory performance

The z score compares the participants deviation from the assigned value with the

standard deviation for proficiency assessment (σpt-log) used as common quality criterion

Measurements that are carried out correctly are assumed to generate results that can be

described by a normal distribution with mean xpt-log and standard deviation σpt-log The z

scores will then be normally distributed with a mean of zero and a standard deviation of

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

1846

10 Only 03 of scores would be expected to fall outside the range -30 lt z lt 30 and

only 5 would be expected to fall outside the range -20 le z le 20 These percentages

may change when the true interlaboratory variability deviates from the agreed standard

deviation which was set to 010 It is unlikely that unacceptable z scores will occur by

chance when no real problem exists rather it is likely that there is an identifiable cause

for any anomaly when an unsatisfactory performance expressed as a z score is obtained

The ζ score states whether the laboratorys result agrees with the assigned value within

the respective measurement uncertainty The denominator is the combined uncertainty of

the assigned value [u(xpt-log)] and the measurement uncertainty as stated by the

laboratory [u(xi-log)] The ζ score includes all parts of a measurement result namely the

expected value (assigned value) its measurement uncertainty in the unit of the result as

well as the uncertainty of the reported values An unsatisfactory ζ score can either be

caused by the presence of a significant bias (inaccurate measurement) or by a non-

realistic estimation of the measurement uncertainty (seriously under-estimated) or by a

combination of both Participants that have obtained a satisfactory z score but an

unsatisfactory ζ score may have underestimated their measurement uncertainty

Participants that have obtained an unsatisfactory z score but a satisfactory ζ score may

have assessed the uncertainty of their result accurately but the result itself does not meet

the performance expected for the PT scheme

More detailed information about measurement uncertainty evaluation can be found in

some international standards and other guidance documents(1718192021)

444 Performance of the laboratories

The performance of the laboratories for GM quantification is primarily evaluated on the

basis of their z scores The ζ scores obtained are providing additional information to the

laboratory regarding the correct estimation of the measurement uncertainty of the result

but should be used as indicative values only

4441 z scores

Table 9 summarises the performance results obtained in this PT based on the z scores

Detailed results per laboratory are reported in Annex 4 Tables A41 and A42 and Figures

A41 and A42

Table 9 Evaluation of laboratory performance for GM event quantification through z scores

Laboratory Performance Test Item 1 Test Item 2

40-3-2 Soybean 40-3-2 Soybean

Number of laboratories with |z| le 20 (satisfactory) 67 78

Number of laboratories with 20 lt |z| lt 30 (questionable) 9 3

Number of laboratories with |z| ge 30 (unsatisfactory) 4a 2 a One additional NRL120 (L88) reported a value ltLOQ which is considered unacceptable

A total of 4 laboratories obtained an unsatisfactory performance expressed as z score for

quantification of event 40-3-2 in T1 (1 NRL882 and 3 non-NRLs) and 2 laboratories for T2

(2 non-NRL) Another 5 NRL882 1 NRL120 and 3 non-NRLs obtained a questionable z

score for T1 and one NRL882 one NRL120 and one non-NRL similarly for T2 In case of

an unsatisfactory performance obtained by an NRL the laboratory will be requested to

perform a root-cause analysis and to communicate the outcome to the EURL GMFF who

will then follow-up with the laboratory

One laboratory (L88) had reported that the 40-3-2 soybean mass fraction in T1 was below

its LOQ of 004 mm While less than X values were not included in the data

evaluation they were compared to the corresponding xpt ndash U(xpt) (after conversion to the

log scale) Since the reported less than X value was lower than the corresponding xpt ndash

U(xpt) the laboratory should have been able to quantify the analyte Therefore the

laboratory statement was considered as unsatisfactory

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

1946

Although the performance scores were calculated on the log-transformed data the

corresponding GM mass fractions on the raw domain which are easier to understand in

daily laboratory practice approximated the following values

For 40-3-2 soybean in T1

Assigned value on the raw domain 080 mm

|z| le 20 lower and upper limits 050 ndash 128 mm

|z| lt 30 lower and upper limits 041 ndash 157 mm

For 40-3-2 soybean in T2

Assigned value on the raw domain 076 mm

|z| le 20 lower and upper limits 048 ndash 122 mm

|z| lt 30 lower and upper limits 039 ndash 150 mm

The general performance of the laboratories for quantification of event 40-3-2 soybean in

both test items was very good This event is one of the older GM events inserted in the

EU Register of authorised events in 1996 and is commonly found in feed samples on the

global market It is therefore not surprising that most laboratories are able to detect this

event and to quantify it accurately In a previous PT in 2014 (CT 0214) which also

included a chicken feed sample containing 40-3-2 soybean the general performance of the

participants was much worse (16 unsatisfactory z scores among 70 results based on an

agreed σpt-log of 020) The latter outcome was due to issues with the extraction of good

quality DNA from the highly processed matrix which contained the 40-3-2 soybean already

before processing (ie the GM soybean was not spiked in) The chicken feed used in the

current PT was from a different origin compared to the one used in CT 0214 although

also in this case the extracted DNA was at least partially degraded In contrast the DNA

from the spiked 40-3-2 soybean was of high-molecular weight (see Section 21) and

therefore easier to amplify during PCR More important is that a considerable fraction of

unprocessed non-GM soybean was added to the T1 mix to increase the total soybean

content the measurements on the DNA extracted from the latter material presumably of

good amplification quality probably contributed mainly to the denominator (ie the taxon-

specific DNA fraction) in the equation to express the GM content The improved

performance of the laboratories participating to the current PT may also be the result of

the increased experience in the extraction of PCR-grade DNA from demanding sample

matrices

Most of the results reported for T2 which was a seed-based matrix and therefore it was

easier to extract good quality DNA were close to the assigned value with its expanded

measurement uncertainty this can be seen in Figure A42 in Annex 4

The participants to this PT were requested to quantify the same GM event in a compound

feed (T1) on the one hand and in a pure soybean material (T2) on the other hand Figure

1 compares the performance of the laboratories to provide acceptable results for both

tasks The horizontal and vertical axes correspond to a z score of 0 for T1 and T2

respectively while the dashed red lines indicate the limits of the satisfactory z scores

(|z|=2) Points in the lower left or upper right quadrant corresponding to L18 and L75

indicate participants who have a systematic bias in the application of the method The

points at the far-left of the x-axis (L26 and L69) represent participants that have had a

problem particularly with T1 but not so much with T2 this may be related to the DNA

extraction part of the workflow which was more challenging for T1 compared to T2 There

are also other participants that may have had similar issues with T1 as more points lie

outside the satisfaction interval on the x-axis compared to the y-axis

For many but not all participants there seem to be a slight tendency for consistent z

scores for T1 and T2 (points along the diagonal line) which gives evidence of participant

bias that affected both test items in a similar way There are also a number of participants

that have obtained a z score close to zero for both test items and which are represented

by points close to where the horizontal and vertical axes cross eg L05 L06 L09 L44

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

2046

L46 L50 L54 L55 L63 L64 L85 L87 and L89 (within |z|lt05 for T1 and T2) These

laboratories 12 NRLs and one non-NRL seem to have the whole analytical method

including DNA extraction and real-time PCR very well under control for different types of

samples The highly proficient non-NRL used digital PCR for both measurements indicating

that this method also seems reliable for GM soybean quantification in different matrices

Figure 1 Youden plot comparing the z scores obtained by the participants for the determination of the 40-3-2 soybean mass fraction in T1 and T2

The dashed red lines indicate the limits of satisfactory performance (|z|le20) for each test item The diagonal line displays the consistency of z scores in T1 and T2

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

2146

4442 ζ scores

Tables A41 and A42 also report the ζ scores obtained by the laboratories as an

informative evaluation for the laboratories of their result in combination with the reported

uncertainty A total of 54 laboratories were given a satisfactory performance when

expressed as ζ score for quantification of 40-3-2 soybean in T1 3 a questionable and 24

an unsatisfactory performance score (N = 81) For 40-3-2 soybean in T2 (N = 83) 62

laboratories performed satisfactorily 5 questionable and 16 unsatisfactorily As explained

in Section 443 a bad ζ score may be due to a result that strongly deviates from the

assigned value (and has therefore also yielded an unsatisfactory performance when

expressed as a z score) or it may indicate an underestimation of the measurement

uncertainty of the result

Figures A41 and A42 (Annex 4) allow verification if the reported measurement

uncertainty bars overlap with the horizontal dashed (black) lines that delineate the

satisfactory interval for the z scores Laboratory L03 provided an expanded measurement

uncertainty of 1520 for T1 and T2 which is either strongly overestimated or corresponds

to a relative uncertainty instead of the requested absolute uncertainty in mm Also L53

has strongly overestimated its measurement uncertainty for the mass fraction of 40-3-2

soybean in T1 and T2 and L76 for T2 only although they both obtained a satisfactory ζ

score because their reported values were sufficiently close to the assigned value From the

same figures in Annex 4 it can easily be seen that several laboratories may have reported

a rather low measurement uncertainty (or no uncertainty at all) and therefore received an

unsatisfactory ζ score eg the NRLs L01 L02 L40 and L78 would have obtained a

satisfactory ζ score for 40-3-2 in T1 if they had reported a realistic measurement

uncertainty (Figure A41) Similarly for 40-3-2 in T2 (Figure A42) the results of the NRLs

L38 L40 L41 L77 L78 and L83 would have been satisfactory if they had reported a

realistic measurement uncertainty Furthermore nearly all laboratories that had not

reported a measurement uncertainty value and thus received an assumed uncertainty of

zero received an unsatisfactory ζ score From the questionnaire it was noted that 32

laboratories (ie more than 13 of all participants) estimated the measurement

uncertainty on the basis of the precision of the analysis replicates only such an approach

may not be sufficient to account for all analytical variability

ISO 135282015(16) suggests to check whether a reported standard uncertainty (with a

coverage factor k=1) is realistic and lies between a minimum and maximum uncertainty

(umin and umax) This allows participants to review their reported uncertainty and evaluate if

the reported uncertainty is counting all relevant components or is over-counting some

components It is unlikely that a participant result will have a smaller standard uncertainty

than the measurement uncertainty of the assigned value so u(xpt) can be used as a lower

limit called umin It is also unlikely that a participant reported standard uncertainty is

larger than the robust standard deviation of the (NRL) results (umax)

As an example Figure 2 compares the relative standard uncertainties (ui) reported by

the participants of this PT for T1 calculated from the expanded uncertainty U and reported

k factor and expressed as percentage of the reported result On the raw data scale umin

and umax for the 40-3-2 soybean mass fraction in T1 correspond to 5 and 28

respectively Therefore a standard measurement uncertainty smaller than 004 mm is

probably underestimated while a standard uncertainty above 023 mm may be

overestimated However these are informative indicators only Measurement uncertainties

below umin or above umax can be valid and in such case the laboratory should check the

result or the uncertainty estimate

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

2246

Figure 2 Reported relative standard uncertainties (ui) for T1

The horizontal blue lines refer to umin and umax as defined in the text Note that the relative

uncertainty of L03 is out of scale and not shown Laboratories that failed to report a measurement uncertainty were given a zero value

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

2346

5 Conclusions

Participants in this PT were required to analyse two test items varying in composition and

complexity but containing the same GM event The analytical tasks resembled the routine

operational analysis tasks of an official control laboratory analysing a food or feed material

for the presence of material derived from containing or consisting of GMOs

The results reported by the participants were analysed and a performance evaluation was

carried out taking into account both the qualitative and the quantitative results reported

A large majority of the participants performed satisfactorily for the tasks in this PT ie the

detection and quantification of the soybean event 40-3-2 in T1 a chicken feed powder

and in T2 a soybean flour All participants who tested for the events were able to identify

the correct event in both test items Regarding quantification four laboratories including

one NRL882 and 3 non-NRLs obtained an unsatisfactory z performance score for the 40-

3-2 soybean measurements in the more difficult feed matrix One NRL120 reported an

unacceptable ltLOQ result for the 40-3-2 soybean mass fraction in T1 but the reported

result for T2 was satisfactory Two of the non-NRL laboratories were also unsatisfactory for

the quantification of the same event in the T2 matrix

It is recommended for several laboratories to re-consider the estimation of their

measurement uncertainty in order to report a more realistic uncertainty and

consequently to obtain a satisfactory ζ performance score

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

2446

Acknowledgements

We acknowledge the support of the members of the Advisory Board for Comparative

Testing (Lotte Hougs Nina Papazova Martin Sandberg and Manuela Schulze) for reviewing

this report and of the Reference Materials Unit of JRC particularly Brigitte Fontenelle for

taking care of the sample dispatching The laboratories listed below are acknowledged for

their participation in this exercise

Body 1 Organisation Department City Country

NRL882

AGES-Institute for Food Safety Vienna Vienna AUSTRIA Umweltbundesamt GmbH Vienna AUSTRIA CRA-W - Walloon Agricultural Research Center Valorization of agric prod Gembloux BELGIUM Institute for Agricultural and Fisheries Research Technology and Food - PI Merelbeke BELGIUM Scientific Institute of Public Health (WIV-ISP) PBB - GMOlab Brussels BELGIUM National Center of Public Health and Analyses GMO Sofia BULGARIA Croatian Institute of Public Health Zagreb CROATIA State General Laboratory GMOs and Allergens Nicosia CYPRUS

Crop Research Institute Prague CZECH REPUBLIC

Danish Veterinary and Food Administration Food Chem and Plant Health Ringsted DENMARK Finnish Customs Laboratory Espoo FINLAND BioGEVES Surgeres FRANCE Service Commun des Laboratoires Illkirch Graffenstad FRANCE Bundesamt fuumlr Verbraucherschutz und Lebensmittelsicherheit

Referat 503 Berlin GERMANY

General Chemical State Laboratory A Chemical Service of Athens Athens GREECE National Food Chain Safety Office Budapest HUNGARY Istituto Zooprofilattico Lazio e Toscana Biotechnology Unit Rome ITALY Institute of Food Safety Animal Health and Environment bdquoBIORrdquo

Riga LATVIA

National Food and Veterinary Risk Assessment Institute

Molecular Biology and GMO Vilnius LITHUANIA

Laboratoire National de Santeacute food control Dudelange LUXEMBOURG RIKILT Wageningen University amp Research Wageningen NETHERLANDS Instytut Zootechniki PIB KLP Szczecin Szczecin POLAND National Veterinary Research Institute Pulawy POLAND Regional Laboratory of Genetically Modified Food Tarnobrzeg POLAND Instituto Nacional de Investigaccedilatildeo Agraacuteria e Veterinaacuteria

UEIS-SAFSV Oeiras PORTUGAL

Institute for Diagnosis and Animal Health Molecular Biology and GMOs Bucharest ROMANIA State Veterinary and Food Institute VFI in Dolny Kubin Dolny Kubin SLOVAKIA Central Control and Testing Institute of Agriculture Bratislava

OMB NRL Bratislava SLOVAKIA

National Institute of Biology Ljubljana SLOVENIA Laboratorio Arbitral Agroalimentario LAA-MAPAMA OGM Madrid SPAIN Centro Nacional De Alimentaciograven (Agencia Espantildea De Consumo Seguridad Alimentaria Y Nutriciograven)

Biotechnology Unit Madrid SPAIN

National Food Agency Uppsala SWEDEN

LGC Teddington UNITED KINGDOM

NRL120

Finnish Food Safety Authority Evira Helsinki FINLAND Thuumlringer Landesamt fuumlr Verbraucherschutz (TLV) Lebensmittelsicherheit Bad Langensalza GERMANY LAVES-Lebensmittel- und Veterinaumlrinstitut BraunschweigHannover

Braunschweig GERMANY

Landesuntersuchungsanstalt fuumlr das Gesundheits- und Veterinaumlrwesen Sachsen

Amtliche Lebensmitteluntersuchung

Dresden GERMANY

BfR Food Safety Berlin GERMANY Landesamt fuumlr Verbraucherschutz Sachsen-Anhalt Fachbereich 3 Halle GERMANY Landesamt fuumlr Landwirtschaft Lebensmittelsicherheit und Fischerei M-V (LALLF MV)

200PCR Rostock GERMANY

Institut fuumlr Hygiene und Umwelt Hamburg Gentechnikuumlberwachungslabor Hamburg GERMANY LUFA Speyer Referat II2 Speyer GERMANY CVUA Freiburg GMO Freiburg GERMANY Bavarian Health and Food Safety Authority (LGL) Oberschleissheim GERMANY LTZ Augustenberg Karlsruhe GERMANY LLBB Berlin GERMANY Landeslabor Schleswig-Holstein Neumuumlnster GERMANY Staatliche Betriebsgesellschaft fuumlr Umwelt und Landwirtschaft

GB 6 Fachbereich 63 Nossen GERMANY

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

2546

0

Body Organisation Department City Country

NRL120 cont

Istituto Superiore di Sanitagrave DSPVSA Rome ITALY

CREA-SCS Sede di Tavazzano Laboratorio

Tavazzano (LO) ITALY

Netherlands Food and Consumer Product Safety Authority (NVWA)

Laboratorium VV Wageningen NETHERLANDS

Plant Breeding and Acclimatization Institute NRI GMO Controlling Laboratory Blonie POLAND

Fera Science Ltd Plants York UNITED KINGDOM

SASA Scottish Government Seed certification Edinburgh UNITED KINGDOM

Non-NRL

FASFC Melle GMO Melle BELGIUM

Laboratoacuterio Nacional Agropecuaacuterio - LANAGROMG Pedro LeopoldoMG BRAZIL

Ministry of Agriculture Livestock and Food Supply Official Laboratory of Goiaacutes Goiania BRAZIL

Laboratory of SGS Bulgaria Ltd Varna BULGARIA

Servicio Agriacutecola y Ganadero Biotechnology Santiago CHILE

Instituto Nacional de Vigilancia de Medicamentos y Alimentos Invima

Laboratorio OGM Bogotaacute COLOMBIA

Croatian Centre for Agriculture Food and Rural Affairs Institute for Seed and Seedlings

Non-NRL Osijek CROATIA

CVUA-OWL Detmold GERMANY

Thuumlringer Landesanstalt fuumlr Landwirtschaft

Jena GERMANY

Biomi Ltd Godollo HUNGARY

ICAR-National Bureau of Plant Genetic Resources Division of Genomic Resources

New Delhi INDIA

IZSLER Brescia ITALY

Istituto Sperimentale Del Piemonte Liguria e Valle DAosta

SC Biotechnologie Torino ITALY

Istituto Zooprofilattico Sperimentale Abruzzo e Molise Hygiene in Food Technology Teramo ITALY

American University of Science and Technology Laboratory Science amp Technology

Ashrafieh-Beirut LEBANON

SENASICA-CNRDOGM Deteccioacuten de OGM Tecaacutemac MEXICO

Bureau of Plant Industry National Plant Quarantine Services Division Post Entry Quarantine Station

Department of Agriculture Los Banos Laguna PHILIPPINES

Laboratorul Central pentru Calitatea Semintelor si a Materialului Saditor Bucuresti

LEDOMG Bucuresti ROMANIA

Institute of Molecular Genetics and Genetic Engineering

Plant Molecular Biology Belgrade SERBIA

SP Laboratorija ad Genetical dpt Becej SERBIA

A Bio Tech Lab Laboratory for biotechnology Sremska Kamenica SERBIA

Agri-Food amp Veterinary Authority of Singapore Veterinary Public Health Labor Singapore SINGAPORE

Federal Food Safety and Veterinary Office FSVO Risk Assessment Division Bern SWITZERLAND

Agroscope Feed Analytics Posieux SWITZERLAND

Ankara Food Control Laboratory Molecular Biology Ankara TURKEY

State Scientific Research Institute of laboratory Diagnostic and Veterynary Sanitary Expertise

Research GMOs Department Kyiv UKRAINE

Ukrmetrteststandart Molecular Biology Kiev UKRAINE

Ukrainian Laboratory of Quality and Safety of Agricultural Products (ULQSAP)

Chabany village UKRAINE

USDA-GIPSA Biotechnology Laboratory Kansas City UNITED STATES

Agricultural Genetics Institute GMO Detection 04 VIETNAM

National Institute for Food Control Quality management Ha Noi VIETNAM

Quality Assurance and Testing Center 3 (QUATEST 3) Microbiology ndash GMO Testing Lab

Bienhoa VIETNAM

1 NRL882 means NRLs designated by their Member State to coordinate the activities of official laboratories for GMO control under Regulation (EC) No 8822004 NRL120 means NRLs nominated under Regulation (EU) No 1202014 to support the EURL GMFF on method validation (and not also NRL882) Non-NRL means official control laboratories from EU or non-EU countries that are not NRLs according to the Regulations mentioned above

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

2646

References

(including those referred to in the Annexes)

1 Regulation (EC) No 18292003 of the European Parliament and of the Council of 22

September 2003 on genetically modified food and feed Off J Eur Union L 268 1-23

2 Regulation (EC) No 8822004 of the European Parliament and of the Council of 29

April 2004 on official controls performed to ensure the verification of compliance with

feed and food law animal health and animal welfare rules Off J Eur Union L 191 1-

52

3 Commission Implementing Regulation (EU) No 1202014 of 7 February 2014

amending Regulation (EC) No 19812006 on detailed rules for the implementation of

Article 32 of Regulation (EC) No 18292003 of the European Parliament and the

Council as regards the Community reference laboratory for genetically modified

organisms Off J Eur Union L 39 46-52

4 Commission Regulation (EU) No 6192011 of 24 June 2011 laying down the methods

of sampling and analysis for the official control of feed as regards presence of

genetically modified material for which an authorisation procedure is pending or the

authorisation of which has expired Off J Eur Union L 166 9-15

5 ISOIEC 170432010 Conformity assessment ndash General requirements for proficiency

testing International Organization for Standardization Geneva Switzerland

6 Dimitrievska B Kortekaas AM Seghers J Leys H Corbisier Ph Trapmann S

(2017) The certification of different mass fractions of MON-Oslash4Oslash32-6 in soya bean

powder Certified Reference Materials ERMreg-BF410ap ERMreg-BF410bp ERMreg-

BF410cp ERMreg-BF410dp and ERMreg-BF410ep EC certification report EUR 28665 EN

ISBN 978-92-79-70107-8

7 Rosa SF Gatto F Angers-Loustau A Petrillo M Kreysa J Querci M (2016)

Development and applicability of a ready-to-use PCR system for GMO screening Food

Chemistry 201 110ndash119

8 Querci M Foti N Bogni B Kluga L Broll H amp Van den Eede G (2009) Real-

time PCR-based ready-to-use multi-target analytical system for GMO detection Food

Anal Methods 2 325-336

9 ISO 170342016 General requirements for the competence of reference material

producers International Organization for Standardization Geneva Switzerland

10 Technical guidance document from the European Union Reference Laboratory for

Genetically Modified Food and Feed on the implementation of Commission Regulation

(EU) No 6192011 (httpgmo-crljrceceuropaeuguidancedocshtm)

11 Corbisier P Barbante A Berben G Broothaerts W De Loose M Emons H

Georgieva Tz Lievens A Mazzara M Papazova N Perri E Sowa S Stebih D Terzi

V Trapmann S (2017) Recommendation for the unit of measurement and the

measuring system to report traceable and comparable results expressing GM content

in accordance with EU legislation JRC Technical report (httpgmo-

crljrceceuropaeuENGLdocsWG-UoM-Final-Reportpdf)

12 Analytical Methods Committee (1989) Robust statistics ndash How not to reject outliers

Part 1 Basic Concepts Analyst 114 1359-1364

13 Analytical Methods Committee (2001) Robust statistics a method for coping with

outliers AMC Technical Brief No 6 April 2001

14 Thompson M Ellison SLR Owen L Mathieson K Powell J Key P Wood R

Damant AP (2006) Scoring in Genetically Modified Organism Proficiency Tests

Based on Log-Transformed Results J AOAC Int 89 232-239

EURL GMFF CT 0217 report

2746

15 Analytical Methods Committee (2004) GMO Proficiency Testing Interpreting z scores

derived from log-transformed data RSC AMC Technical Brief No 18 December

2004

16 ISO 135282015 Statistical methods for use in proficiency testing by interlaboratory

comparison International Organization for Standardization Geneva Switzerland

17 Ellison SLR Williams A (Eds) EurachemCITAC guide Quantifying Uncertainty in

Analytical Measurement Third edition (2012) ISBN 978-0-948926-30-3 Available

from wwweurachemorg

18 ISOIEC Guide 98-32008 (JCGMWG1100) Uncertainty of measurement - Part 3

Guide to the expression of uncertainty in measurement

19 Linsinger T (2010) Comparison of a measurement result with the certified value

Application Note 1 Luxemburg Office for Official Publications of the European

Communities (httpscrmjrceceuropaeu)

20 Trapmann S Burns M Broll H Macarthur R Wood R Zel J (2009) Guidance

Document on Measurement Uncertainty for GMO Testing Laboratories Luxembourg

Office for Official Publications of the European Communities EUR 22756 EN2-2009

21 Corbisier P Zobell O Trapmann S Auclair G and Emons H (2014) Training

Manual on GMO Quantification Proper Calibration and Estimation of Measurement

Uncertainty Luxembourg Publications Office of the European Union

(doi10278710085)

22 Thompson M Wood R (1993) The international harmonized protocol for the

proficiency testing of (chemical) analytical laboratories J AOAC Int 76 926-940

23 Powell J Owen L (2002) Reliability of Food Measurements The Application of

Proficiency Testing to GMO Analysis Accred Qual Ass 7 392-402

24 Linsinger TPJ van der Veen AMH Gawlik BM Pauwels J Lamberty A

(2004) Planning and combining of isochronous stability studies of CRMs Accred Qual

Assur 9 464-472

EURL GMFF CT 0217 report

2846

Annexes

Annex 1 Homogeneity and stability of test items

A11 Homogeneity of test items

The homogeneity of T2 was confirmed during the certification of this CRM

The assessment of the homogeneity(16) of T1 was performed after the test item had been

packed in its final form and before distribution to participants using the following

acceptance criterion

pts s 30 (A11)

Where ss is the between-test item standard deviation as determined by a 1-way random

effects ANOVA(22) and σpt is the standard deviation for comparative testing The value of

σpt the target standard deviation for proficiency assessment was based upon the

experience acquired in previous PTs and set to 010 on the log domain(23)

If the criterion according to A11 is met (ie ss le 0030) the between-test item standard

deviation contributes no more than about 10 to the standard deviation for comparative

testing

The repeatability of the test method is the square root of the mean sum of squares within-

test items MSwithin The relative between-test item standard deviation ssrel is given by

100

y

n

MSMS

s

withinbetween

rels (A12)

where MSbetween is the mean sum of squares between test items

MSwithin is the mean sum of squares within test items

n is the number of replicates for each sample

y is the mean of the homogeneity data

If MSwithin gt MSbetween then

100

1

24

y

nNn

ityrepeatabil

us bbrels (A13)

where ubb is the maximum uncertainty contribution that can be obtained by the hidden

heterogeneity of the material

Seven bottles (N = 7) were randomly selected and analysed in five replicates (n = 5) The

between-test item standard deviation was 0018 mm The criterion described in

formula (A11) was fulfilled (0018 lt 0030) indicating that T1 was adequately

homogeneous

EURL GMFF CT 0217 report

2946

A12 Stability of test items

For T1 an isochronous short-term stability study(24) involving two test samples with three

replicates each (N = 2 n = 3) was conducted over two and four weeks at +4 degC +18 degC

and +60 degC The 40-3-2 soybean mass fraction was measured by qPCR The

measurements were performed under intermediate precision conditions with respect to the

PCR plates

The results did not reveal any influence of time or storage at +4 degC or +18 degC on the

stability of the test item (compared to storage at -70 degC) with regard to the soybean event

40-3-2 mass fraction Even at 60 degC no significant trend was measured although the

extracted DNA was more fragmented as seen by agarose gel electrophoresis

The test items were shipped at ambient temperature

The stability of T1 during the period covered by the PT (approximately 5 months between

production of the test item and the deadline for results reporting) was tested by analysing

simultaneously on one PCR plate two units (N = 2 n = 3) stored either at the normal

storage temperature (4 degC) or at a reference temperature (-70 degC) The evaluation was

based on the results ratio between samples stored at 4 degC and -70 degC The data were

evaluated against storage time and regression lines were calculated The slopes of the

regression lines were tested for statistical significance (lossincrease due to storage) No

significant trend was detected at a 95 confidence level The T1 material can therefore

be stored at 4 degC and was stable during the period covered by this CT

The stability of T2 was ensured as part of the post-certification stability monitoring of

ERM-BF410p Measurements were performed simultaneously on one PCR plate as

described for T1 on units stored at the normal storage temperature (4 degC) and at a

reference temperature (-70 degC) No significant trend was detected at a 95 confidence

level The T2 material can therefore be stored at 4 degC and was stable during the period

covered by this CT

EURL GMFF CT 0217 report

3046

Annex 2 Questionnaire data

The results received from 82 laboratories were exported from the EUSurvey

Questionnaire on CT 0217 analysis and are tabulated below Multiple answers were

allowed for all questions except for the questions on the calibrant used The results of the

open questions were manually analysed and reported Answers to the questions on GM

events that were not to be quantified in the test items are not shown

Select the group to which your organisation belongs Note 882 and 120 refer to EU Regulations 8822004 and 1202014

resp select NRL120 if your organisation is ONLY listed under Regulation 1202014 select non-NRL if your organisation is

not an NRL under either EU Regulation

Answers Ratio

NRL882 33 402

NRL120 20 244

Non-NRL 29 354

No Answer 0 0

T1 Please select the option that applies and proceed with the questionnaire

Answers Ratio

T1 was not analysed go to Q1 2 244

T1 was analysed go to Q2 80 976

No Answer 0 0

T1 1 Why did you not analyse test item 1

Answers Ratio

a) The sample matrix is out of the scope of our laboratory 1 122

b) The methods are not validated in our laboratory 0 0

c) We could not obtain sufficient good quality DNA suitable for further analysis 0 0

d) Reference material primers probes or other reagents were not available (in time) 0 0

e) We tried but our analysis failed 0 0

f) Other practical constraints (instrument broken no personnel etc) 1 122

g) Other reason 0 0

No Answer 80 976

T1 2 Select the DNA extraction method used for T1

Answers Ratio

CTAB 41 50

NucleoSpin Food 10 122

NucleoSpin Plant 3 366

GeneSpin 4 488

Promega Wizard 3 366

DNeasy Plant 3 366

DNeasy Mericon Food 5 61

Biotecon Foodproof 5 61

SDS 4 488

Fast ID Genomic DNA 2 244

Maxwell 16 Plant DNA 0 0

Maxwell 16 Food Feed Seed 5 61

Generon Ion Force 1 122

Other 2 244

No Answer 2 244

T1 3 Select any additional DNA purification method used for T1

Answers Ratio

No additional clean-up 51 622

Additional ethanol precipitation 9 110

Eurofins DNAExtractor cleaning column 3 366

Promega Wizard DNA clean-up resin 6 732

Qiagen QIAQuick 4 488

Qiagen Genomic-Tip 20G 1 122

Other method (no need to specify) 7 854

No Answer 2 244

EURL GMFF CT 0217 report

3146

T1 4 Indicate the number of replicate DNA extractions used to obtain the results

Answers Ratio

1 0 0

2 52 634

3 9 110

4 10 122

5 2 244

6 6 732

gt6 1 122

No Answer 2 244

T1 5 Select the approach(es) used to show absence of PCR inhibition

Answers Ratio None (no inhibition was suspected based on experience) 6 732

We check that the optical density ratios (OD260280 260230) are acceptable 40 488

We verify that the amplification curves look normal 24 293

We run two dilutions and verify if the delta Cq or GM are as expected 31 378

We run three or four dilutions and verify if the delta Cq or GM are as expected 13 158

We perform a PCR inhibition run with a reference gene before analysis 3 or 4 dilutions linear regression extrapolation of Cq for undiluted extract compare this to the measured Cq

16 195

We add an internal positive control to the reactions and check the Cq 15 183

Other 1 122

No Answer 2 244

T1 6 Select the option applicable to your analysis for 40-3-2 soybean in T1

Answers Ratio

Not tested 0 0

Detected but not quantified 3 366

Detected and quantified please fill in Q6a-6h 77 939

Found absent 0 0

No Answer 2 244

T1 7 Select the option applicable to your analysis for 68416 soybean in T1

Answers Ratio

Not tested 14 171

Detected but not quantified 2 244

Detected and quantified please fill in Q7a-7h 0 0

Found absent 64 781

No Answer 2 244

T1 8 Select the option applicable to your analysis for MON89788 soybean in T1

Answers Ratio

Not tested 6 732

Detected but not quantified 33 402

Detected and quantified please fill in Q8a-8h 15 183

Found absent 26 317

No Answer 2 244

T1 9 If applicable why did you not test or quantify all GM events in T1

Answers Ratio

a) Not applicable all GM events listed were tested and all those detected were quantified 45 549

b) The event-specific detection method is not validated in our laboratory 7 854

c) Reference material primers probes or other reagents were not available (in time) 9 110

d) The result obtained was below the LODLOQ 18 220

e) Practical constraints (instrument broken no personnel etc) 0 0

f) Other reason 9 110

No Answer 2 244

T1 6a Soybean 40-3-2 Which quantification approach was used

Answers Ratio

Standard curve method (2 calibration curves) 68 829

Delta Cq method (one calibration curve) 9 110

Digital PCR (no calibration curve) 2 244

EURL GMFF CT 0217 report

3246

No Answer 5 61

T1 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 71 866

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 5 61

T1 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T1

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 65 793

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 5 61

Other please specify below 1 122

No Answer 5 61

Specify the reference target(s) used (if different from above)

SOJA LEKTIN 80 bp - Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

T1 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

59 720

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

13 159

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 5 61

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T1 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 9 110

100 ng 18 220

50 ng 14 171

25 ng 6 732

15 ng 0 0

lt10 ng 0 0

No Answer 5 61

EURL GMFF CT 0217 report

3346

T1 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 3 002 4 003 1 004 6 005 7 006 2 008 4 009 7 01 37 019 1 02 1 03 1 035 1 04 1 13 1

T1 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 20 244

Determined during the in-house validation of the method 42 512

Taken from the EURL GMFF validation report 17 207

By another approach please explain below 2 244

No Answer 5 61

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

T1 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2 soybean

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 27 329

Known uncertainty of the standard method 9 110

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 6 732

No Answer 5 61

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx

Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

we used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

MON89788 soybean was detected but not quantified because results for this event were lt LOQ

We saw in the qPCR for MON89788 Ct-values at 37 - 39 but lt LOD (lt 002 )

Additional tests with one extract performed with Nucleospin Food Kit and with one extract performed with CTAB+Nucleospin Food Kit showed no significant differences in event detection and GTS 40-3-2 quantification (examined with digital PCR)

T19 not all detected GM were asked to be quantified

For T1 trace amounts of MON89788 were detected either in qualitative screening as well as in quantification The givencalculated GM (mm) of 002 is well below the known and validated LOQ of the method which is 01 We detected a low level of MON89788 but as this quantified at 0006 lower than the threshold for reporting we took this as a negative result

According course material of JRC GMO Quantification Proper calibration and Estimation of Measurement Uncertainty

MU is based on intermediate precision data

EURL GMFF CT 0217 report

3446

The sample was strongly inhibited MON89788 was detected under LOQ

In T1 the measured GM for MON89788 of 001 is below the determined LOQ (02) therefore no MU was reported

MON89788 was detected in traces at the LOD

MON89788 SOY result obtained below the LOQ but copy numbers of MON89788 was above LOD

T1 was inhibited A 10x dilution of the DNA was needed to get a reasonable result

DNA extraction of T1 has conducted using DNA extraction kits SureFood Prep Basic (S1052) and SureFood Prep Advanced (S1053) R-Biopharm AG Identification of Soya GM-lines was conducted using diagnostic kits SureFood GMO ID Roundup Ready Soya (S2030) and SureFood GMO ID RR2Y (S2034) Quantification kits were SureFood GMO Quant Roundup Ready Soya (S2014) SureFood GMO Quant 35S soya (S2028) and SureFood GMO Quant RR2Y Soya (S2029) R-Biopharm AG MON89788 soybean was detected in sample T1 but only in some cases - 2 PCR replicates were always positive out of 4 after several repetitions

Remark (precision) in relation to T18 and T19 MON89788 soybean was detected in the sample but not quantified because of lt LOQ

T2 Please select the option that applies and proceed with the questionnaire

Answers Ratio

T2 was not analysed go to Q1 0 0

T2 was analysed go to Q2 82 100

No Answer 0 0

T2 1 Why did you not analyse test item 2

Answers Ratio

a) The sample matrix is out of the scope of our laboratory 0 0

b) The methods are not validated in our laboratory 0 0

c) We could not obtain sufficient good quality DNA suitable for further analysis 0 0

d) Reference material primers probes or other reagents were not available (in time) 0 0

e) We tried but our analysis failed 0 0

f) Other practical constraints (instrument broken no personnel etc) 0 0

g) Other reason 0 0

No Answer 82 100

T2 2 Select the DNA extraction method used for T2

Answers Ratio

CTAB 40 488

NucleoSpin Food 11 134

NucleoSpin Plant 3 366

GeneSpin 4 488

Promega Wizard 3 366

DNeasy Plant 3 366

DNeasy Mericon Food 4 488

Biotecon Foodproof 5 61

SDS 4 488

Fast ID Genomic DNA 3 366

Maxwell 16 Plant DNA 0 0

Maxwell 16 Food Feed Seed 4 488

Generon Ion Force 1 122

Other 2 244

No Answer 0 0

T2 3 Select any additional DNA purification method used for T2

Answers Ratio

No additional clean-up 52 634

Additional ethanol precipitation 9 110

Eurofins DNAExtractor cleaning column 3 366

Promega Wizard DNA clean-up resin 6 732

Qiagen QIAQuick 5 61

Qiagen Genomic-Tip 20G 1 122

Other method (no need to specify) 8 976

No Answer 0 0

EURL GMFF CT 0217 report

3546

T2 4 Indicate the number of replicate DNA extractions used to obtain the results

Answers Ratio

1 0 0

2 56 683

3 9 110

4 10 122

5 2 244

6 4 488

gt6 1 122

No Answer 0 0

T2 5 Select the approach(es) used to show absence of PCR inhibition

Answers Ratio

None (no inhibition was suspected based on experience) 7 854

We run two dilutions and verify if the delta Cq or GM are as expected 32 390

We run three or four dilutions and verify if the delta Cq or GM are as expected 12 146

We perform a PCR inhibition run with a reference gene before analysis 3 or 4 dilutions linear regression extrapolation of Cq of undiluted extract compare this to the measured Cq

15 183

We add an internal positive control to the reactions and check the Cq 16 195

We verify that the amplification curves look normal 25 305

We check that the optical density ratios (OD260280 260230) are acceptable 36 439

Other 1 122

No Answer 0 0

T2 6 Select the option applicable to your analysis for 40-3-2 soybean in T2

Answers Ratio

Not tested 0 0

Detected but not quantified 2 244

Detected and quantified please fill in Q6a-6h 80 976

Found absent 0 0

No Answer 0 0

T2 7 Select the option applicable to your analysis for 68416 soybean in T2

Answers Ratio

Not tested 15 183

Detected but not quantified 1 122

Detected and quantified please fill in Q7a-7h 0 0

Found absent 66 805

No Answer 0 0

T2 8 Select the option applicable to your analysis for MON89788 soybean in T2

Answers Ratio

Not tested 5 61

Detected but not quantified 6 732

Detected and quantified please fill in Q8a-8h 3 366

Found absent 68 829

No Answer 0 0

T2 9 If applicable why did you not test or quantify all GM events in T2

Answers Ratio

a) Not applicable all GM events listed were tested and all those detected were quantified 61 744

b) The event-specific detection method is not validated in our laboratory 6 732

c) Reference material primers probes or other reagents were not available (in time) 10 122

d) The result obtained was below the LODLOQ 5 61

e) Practical constraints (instrument broken no personnel etc) 0 0

f) Other reason 2 244

No Answer 0 0

T2 6a 40-3-2 Which quantification approach was used

Answers Ratio

Standard curve method (2 calibration curves) 71 866

Delta Cq method (one calibration curve) 9 110

Digital PCR (no calibration curve) 2 244

EURL GMFF CT 0217 report

3646

No Answer 2 244

T2 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 74 902

CRM from IRMM certified in GM copy number ratio (plasmid CRM) 0 0

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 2 244

T2 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T2

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 63 768

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 6 732

Other please specify below 2 244

No Answer 5 61

Specify the reference target(s) used (if different from above)

Lectin - 74bp

lec 74bp

Lectine

SOJA LEKTIN 80 bp Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

Terry C F Harris N Event-specific detection of Roundup Ready soya using two different real time PCR detection chemistries Eur Food Res Technol (2001) 213425-431

T2 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

61 744

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

14 171

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 2 244

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T2 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 7 854

100 ng 21 256

EURL GMFF CT 0217 report

3746

50 ng 17 207

25 ng 7 854

15 ng 2 244

lt10 ng 0 0

No Answer 2 244

T2 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 5 002 6 003 1 004 7 005 7 006 2 007 1 008 2 009 8 01 39 026 1 028 1

T2 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 21 256

Determined during the in-house validation of the method 43 524

Taken from the EURL GMFF validation report 19 232

By another approach please explain below 3 366

No Answer 2 244

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample

Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit

T2 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 30 366

Known uncertainty of the standard method 8 976

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 7 854

No Answer 2 244

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

We used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

30 of the quantification result

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

Event 68416 was not tested due to non-availability of Positive control and reference material

DNA extraction of T2 has conducted using DNA extraction kits SureFood Prep Basic etc (see T1)

The NucleoSpin Food + NucleoSpin gDNA Clean-up was also used for DNA extraction from T1 and T2 samples The results for 40-3-2 were 25 - 30 lower than CTAB results T1 050 +- 018 (ww) T2 058 +- 020 (ww)

AOCS 0906-B as calibrant for MON89788 in T2 also

EURL GMFF CT 0217 report

3846

Annex 3 Qualitative performance of the participants

Table A31 Performance of the participants for the qualitative identification of GM events in CT-0217

The correct answer (yes) is that the GM event has been detected in the test item

Lab 40-3-2 detected in T1 40-3-2 detected in T2 Lab 40-3-2 detected in T1 40-3-2 detected in T2

L01 yes yes L45 yes yes

L02 yes yes L46 yes yes

L03 yes yes L47 yes yes

L04 yes yes L48 yes yes

L05 yes yes L49 yes yes

L06 yes yes L50 yes yes

L07 yes yes L51 yes yes

L08 yes yes L52 yes yes

L09 yes yes L53 yes yes

L10 yes Yes L54 yes yes

L11 yes yes L55 yes yes

L12 yes yes L56 yes yes

L13 yes yes L57 yes yes

L14 yes yes L58 yes yes

L15 yes yes L59 yes yes

L16 yes yes L60 yes yes

L17 yes yes L61 yes yes

L18 yes yes L62 yes yes

L19 yes yes L63 yes yes

L20 yes yes L66 yes yes

L22 yes yes L68 yes yes

L23 yes yes L69 yes yes

L24 yes yes L70 yes yes

L25 yes yes L71 yes yes

L26 yes yes L72 yes yes

L27 yes yes L73 T1 was not analysed yes

L29 yes yes L74 yes yes

L30 yes yes L75 yes yes

L31 yes yes L76 yes yes

L32 yes yes L77 yes yes

L33 yes yes L78 yes yes

L34 yes yes L79 yes yes

L35 yes yes L80 yes yes

L36 yes yes L81 yes yes

L37 yes yes L82 yes yes

L38 yes yes L83 yes yes

L39 yes yes L84 yes yes

L40 yes yes L85 yes yes

L41 yes yes L86 yes yes

L42 yes yes L87 yes yes

L43 T1 was not analysed yes L88 yes yes

L44 yes yes L89 yes yes

Although the questionnaire was not returned by the participant the presence of the event was inferred from the quantitative result reported

L10 did not fill in the questionnaire but reported a larger than 01 result for T1 and T2

EURL GMFF CT 0217 report

3946

Annex 4 Participants quantitative performance

The z and ζ scores of all laboratories are reported in Tables A41 and A42 and in Figures

A41 and A42 for 40-3-2 soybean in T1 and T2 respectively For consistency the

reported results are shown with two decimals the performance scores with one decimal

The z scores are displayed in green (satisfactory) orange (questionable) or red cells

(unsatisfactory)

Table A41 Quantitative results (in mm ) and performance scores of participants of CT-0217 for 40-3-2 soybean in chicken feed (T1)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 064 008 200 -019 004 003 -10 -28

L02 NRL882 057 002 200 -024 001 001 -15 -67

L03 Non-NRL 073 1520 200 -014 760 452 -04 00

L04 Non-NRL 075

-012 000 000 -03 -13

L05 NRL120 073 007 257 -014 003 002 -04 -15

L06 NRL882 075 021 200 -012 011 006 -03 -04

L07 NRL882 065 013 200 -019 007 004 -09 -19

L08 NRL120 062 023 200 -021 012 008 -11 -13

L09 NRL882 081 027 200 -009 014 007 01 01

L12 Non-NRL 100 051 200 000 026 011 10 09

L13 NRL882 094 019 173 -003 011 005 07 13

L14 Non-NRL 063

-020 000 000 -10 -50

L15 NRL882 110 011 200 004 005 002 14 46

L16 Non-NRL 117

007 000 000 17 80

L17 NRL882 072 010 200 -014 005 003 -05 -12

L18 Non-NRL 040

-040 000 000 -30 -145

L19 Non-NRL 115 063 173 006 036 014 16 11

L20 NRL120 108 032 200 003 016 006 13 19

L22 Non-NRL 051 030 223 -029 013 011 -20 -17

L23 Non-NRL 047 009 200 -033 005 004 -23 -50

L24 NRL882 072 019 200 -014 010 006 -05 -07

L25 Non-NRL 120

008 000 000 18 85

L26 NRL882 023 009 200 -064 005 008 -54 -62

L27 Non-NRL 098 015 200 -001 008 003 09 23

L29 NRL120 090 008 173 -005 005 002 05 17

L30 NRL882 045 013 200 -035 007 006 -25 -38

L31 NRL882 080 038 223 -010 017 009 00 00

L32 Non-NRL 076 039 200 -012 020 011 -02 -02

L33 NRL-120 057 014 200 -024 007 005 -15 -26

L34 NRL120 061 016 200 -021 008 006 -12 -19

L35 NRL882 097 027 200 -001 014 006 08 13

L36 Non-NRL 090 030 200 -005 015 007 05 07

L37 NRL120 065 026 200 -019 013 009 -09 -10

L38 NRL120 093 014 237 -003 006 003 07 19

L39 Non-NRL 112

005 000 000 15 71

L40 NRL882 053 008 200 -028 004 003 -18 -46

L41 NRL120 080 003 200 -010 002 001 00 00

L42 NRL882 104 031 200 002 016 006 11 17

L44 NRL120 084 007 257 -008 003 001 02 09

L45 NRL882 065 018 200 -019 009 006 -09 -14

L46 NRL882 081 018 200 -009 009 005 01 01

L47 NRL882 141 027 173 015 016 005 25 47

L48 NRL882 059 028 200 -023 014 010 -13 -13

L49 Non-NRL 089 031 200 -005 016 008 05 06

L50 NRL120 073 005 200 -014 003 001 -04 -15

L51 Non-NRL 077 049 200 -011 025 014 -02 -01

L52 Non-NRL 091 012 200 -004 006 003 06 16

L53 Non-NRL 043 081 200 -037 041 041 -27 -07

L54 NRL882 074 035 200 -013 018 010 -03 -03

L55 Non-NRL 076 019 173 -012 011 006 -02 -03

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

2946

A12 Stability of test items

For T1 an isochronous short-term stability study(24) involving two test samples with three

replicates each (N = 2 n = 3) was conducted over two and four weeks at +4 degC +18 degC

and +60 degC The 40-3-2 soybean mass fraction was measured by qPCR The

measurements were performed under intermediate precision conditions with respect to the

PCR plates

The results did not reveal any influence of time or storage at +4 degC or +18 degC on the

stability of the test item (compared to storage at -70 degC) with regard to the soybean event

40-3-2 mass fraction Even at 60 degC no significant trend was measured although the

extracted DNA was more fragmented as seen by agarose gel electrophoresis

The test items were shipped at ambient temperature

The stability of T1 during the period covered by the PT (approximately 5 months between

production of the test item and the deadline for results reporting) was tested by analysing

simultaneously on one PCR plate two units (N = 2 n = 3) stored either at the normal

storage temperature (4 degC) or at a reference temperature (-70 degC) The evaluation was

based on the results ratio between samples stored at 4 degC and -70 degC The data were

evaluated against storage time and regression lines were calculated The slopes of the

regression lines were tested for statistical significance (lossincrease due to storage) No

significant trend was detected at a 95 confidence level The T1 material can therefore

be stored at 4 degC and was stable during the period covered by this CT

The stability of T2 was ensured as part of the post-certification stability monitoring of

ERM-BF410p Measurements were performed simultaneously on one PCR plate as

described for T1 on units stored at the normal storage temperature (4 degC) and at a

reference temperature (-70 degC) No significant trend was detected at a 95 confidence

level The T2 material can therefore be stored at 4 degC and was stable during the period

covered by this CT

EURL GMFF CT 0217 report

3046

Annex 2 Questionnaire data

The results received from 82 laboratories were exported from the EUSurvey

Questionnaire on CT 0217 analysis and are tabulated below Multiple answers were

allowed for all questions except for the questions on the calibrant used The results of the

open questions were manually analysed and reported Answers to the questions on GM

events that were not to be quantified in the test items are not shown

Select the group to which your organisation belongs Note 882 and 120 refer to EU Regulations 8822004 and 1202014

resp select NRL120 if your organisation is ONLY listed under Regulation 1202014 select non-NRL if your organisation is

not an NRL under either EU Regulation

Answers Ratio

NRL882 33 402

NRL120 20 244

Non-NRL 29 354

No Answer 0 0

T1 Please select the option that applies and proceed with the questionnaire

Answers Ratio

T1 was not analysed go to Q1 2 244

T1 was analysed go to Q2 80 976

No Answer 0 0

T1 1 Why did you not analyse test item 1

Answers Ratio

a) The sample matrix is out of the scope of our laboratory 1 122

b) The methods are not validated in our laboratory 0 0

c) We could not obtain sufficient good quality DNA suitable for further analysis 0 0

d) Reference material primers probes or other reagents were not available (in time) 0 0

e) We tried but our analysis failed 0 0

f) Other practical constraints (instrument broken no personnel etc) 1 122

g) Other reason 0 0

No Answer 80 976

T1 2 Select the DNA extraction method used for T1

Answers Ratio

CTAB 41 50

NucleoSpin Food 10 122

NucleoSpin Plant 3 366

GeneSpin 4 488

Promega Wizard 3 366

DNeasy Plant 3 366

DNeasy Mericon Food 5 61

Biotecon Foodproof 5 61

SDS 4 488

Fast ID Genomic DNA 2 244

Maxwell 16 Plant DNA 0 0

Maxwell 16 Food Feed Seed 5 61

Generon Ion Force 1 122

Other 2 244

No Answer 2 244

T1 3 Select any additional DNA purification method used for T1

Answers Ratio

No additional clean-up 51 622

Additional ethanol precipitation 9 110

Eurofins DNAExtractor cleaning column 3 366

Promega Wizard DNA clean-up resin 6 732

Qiagen QIAQuick 4 488

Qiagen Genomic-Tip 20G 1 122

Other method (no need to specify) 7 854

No Answer 2 244

EURL GMFF CT 0217 report

3146

T1 4 Indicate the number of replicate DNA extractions used to obtain the results

Answers Ratio

1 0 0

2 52 634

3 9 110

4 10 122

5 2 244

6 6 732

gt6 1 122

No Answer 2 244

T1 5 Select the approach(es) used to show absence of PCR inhibition

Answers Ratio None (no inhibition was suspected based on experience) 6 732

We check that the optical density ratios (OD260280 260230) are acceptable 40 488

We verify that the amplification curves look normal 24 293

We run two dilutions and verify if the delta Cq or GM are as expected 31 378

We run three or four dilutions and verify if the delta Cq or GM are as expected 13 158

We perform a PCR inhibition run with a reference gene before analysis 3 or 4 dilutions linear regression extrapolation of Cq for undiluted extract compare this to the measured Cq

16 195

We add an internal positive control to the reactions and check the Cq 15 183

Other 1 122

No Answer 2 244

T1 6 Select the option applicable to your analysis for 40-3-2 soybean in T1

Answers Ratio

Not tested 0 0

Detected but not quantified 3 366

Detected and quantified please fill in Q6a-6h 77 939

Found absent 0 0

No Answer 2 244

T1 7 Select the option applicable to your analysis for 68416 soybean in T1

Answers Ratio

Not tested 14 171

Detected but not quantified 2 244

Detected and quantified please fill in Q7a-7h 0 0

Found absent 64 781

No Answer 2 244

T1 8 Select the option applicable to your analysis for MON89788 soybean in T1

Answers Ratio

Not tested 6 732

Detected but not quantified 33 402

Detected and quantified please fill in Q8a-8h 15 183

Found absent 26 317

No Answer 2 244

T1 9 If applicable why did you not test or quantify all GM events in T1

Answers Ratio

a) Not applicable all GM events listed were tested and all those detected were quantified 45 549

b) The event-specific detection method is not validated in our laboratory 7 854

c) Reference material primers probes or other reagents were not available (in time) 9 110

d) The result obtained was below the LODLOQ 18 220

e) Practical constraints (instrument broken no personnel etc) 0 0

f) Other reason 9 110

No Answer 2 244

T1 6a Soybean 40-3-2 Which quantification approach was used

Answers Ratio

Standard curve method (2 calibration curves) 68 829

Delta Cq method (one calibration curve) 9 110

Digital PCR (no calibration curve) 2 244

EURL GMFF CT 0217 report

3246

No Answer 5 61

T1 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 71 866

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 5 61

T1 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T1

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 65 793

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 5 61

Other please specify below 1 122

No Answer 5 61

Specify the reference target(s) used (if different from above)

SOJA LEKTIN 80 bp - Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

T1 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

59 720

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

13 159

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 5 61

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T1 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 9 110

100 ng 18 220

50 ng 14 171

25 ng 6 732

15 ng 0 0

lt10 ng 0 0

No Answer 5 61

EURL GMFF CT 0217 report

3346

T1 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 3 002 4 003 1 004 6 005 7 006 2 008 4 009 7 01 37 019 1 02 1 03 1 035 1 04 1 13 1

T1 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 20 244

Determined during the in-house validation of the method 42 512

Taken from the EURL GMFF validation report 17 207

By another approach please explain below 2 244

No Answer 5 61

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

T1 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2 soybean

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 27 329

Known uncertainty of the standard method 9 110

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 6 732

No Answer 5 61

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx

Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

we used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

MON89788 soybean was detected but not quantified because results for this event were lt LOQ

We saw in the qPCR for MON89788 Ct-values at 37 - 39 but lt LOD (lt 002 )

Additional tests with one extract performed with Nucleospin Food Kit and with one extract performed with CTAB+Nucleospin Food Kit showed no significant differences in event detection and GTS 40-3-2 quantification (examined with digital PCR)

T19 not all detected GM were asked to be quantified

For T1 trace amounts of MON89788 were detected either in qualitative screening as well as in quantification The givencalculated GM (mm) of 002 is well below the known and validated LOQ of the method which is 01 We detected a low level of MON89788 but as this quantified at 0006 lower than the threshold for reporting we took this as a negative result

According course material of JRC GMO Quantification Proper calibration and Estimation of Measurement Uncertainty

MU is based on intermediate precision data

EURL GMFF CT 0217 report

3446

The sample was strongly inhibited MON89788 was detected under LOQ

In T1 the measured GM for MON89788 of 001 is below the determined LOQ (02) therefore no MU was reported

MON89788 was detected in traces at the LOD

MON89788 SOY result obtained below the LOQ but copy numbers of MON89788 was above LOD

T1 was inhibited A 10x dilution of the DNA was needed to get a reasonable result

DNA extraction of T1 has conducted using DNA extraction kits SureFood Prep Basic (S1052) and SureFood Prep Advanced (S1053) R-Biopharm AG Identification of Soya GM-lines was conducted using diagnostic kits SureFood GMO ID Roundup Ready Soya (S2030) and SureFood GMO ID RR2Y (S2034) Quantification kits were SureFood GMO Quant Roundup Ready Soya (S2014) SureFood GMO Quant 35S soya (S2028) and SureFood GMO Quant RR2Y Soya (S2029) R-Biopharm AG MON89788 soybean was detected in sample T1 but only in some cases - 2 PCR replicates were always positive out of 4 after several repetitions

Remark (precision) in relation to T18 and T19 MON89788 soybean was detected in the sample but not quantified because of lt LOQ

T2 Please select the option that applies and proceed with the questionnaire

Answers Ratio

T2 was not analysed go to Q1 0 0

T2 was analysed go to Q2 82 100

No Answer 0 0

T2 1 Why did you not analyse test item 2

Answers Ratio

a) The sample matrix is out of the scope of our laboratory 0 0

b) The methods are not validated in our laboratory 0 0

c) We could not obtain sufficient good quality DNA suitable for further analysis 0 0

d) Reference material primers probes or other reagents were not available (in time) 0 0

e) We tried but our analysis failed 0 0

f) Other practical constraints (instrument broken no personnel etc) 0 0

g) Other reason 0 0

No Answer 82 100

T2 2 Select the DNA extraction method used for T2

Answers Ratio

CTAB 40 488

NucleoSpin Food 11 134

NucleoSpin Plant 3 366

GeneSpin 4 488

Promega Wizard 3 366

DNeasy Plant 3 366

DNeasy Mericon Food 4 488

Biotecon Foodproof 5 61

SDS 4 488

Fast ID Genomic DNA 3 366

Maxwell 16 Plant DNA 0 0

Maxwell 16 Food Feed Seed 4 488

Generon Ion Force 1 122

Other 2 244

No Answer 0 0

T2 3 Select any additional DNA purification method used for T2

Answers Ratio

No additional clean-up 52 634

Additional ethanol precipitation 9 110

Eurofins DNAExtractor cleaning column 3 366

Promega Wizard DNA clean-up resin 6 732

Qiagen QIAQuick 5 61

Qiagen Genomic-Tip 20G 1 122

Other method (no need to specify) 8 976

No Answer 0 0

EURL GMFF CT 0217 report

3546

T2 4 Indicate the number of replicate DNA extractions used to obtain the results

Answers Ratio

1 0 0

2 56 683

3 9 110

4 10 122

5 2 244

6 4 488

gt6 1 122

No Answer 0 0

T2 5 Select the approach(es) used to show absence of PCR inhibition

Answers Ratio

None (no inhibition was suspected based on experience) 7 854

We run two dilutions and verify if the delta Cq or GM are as expected 32 390

We run three or four dilutions and verify if the delta Cq or GM are as expected 12 146

We perform a PCR inhibition run with a reference gene before analysis 3 or 4 dilutions linear regression extrapolation of Cq of undiluted extract compare this to the measured Cq

15 183

We add an internal positive control to the reactions and check the Cq 16 195

We verify that the amplification curves look normal 25 305

We check that the optical density ratios (OD260280 260230) are acceptable 36 439

Other 1 122

No Answer 0 0

T2 6 Select the option applicable to your analysis for 40-3-2 soybean in T2

Answers Ratio

Not tested 0 0

Detected but not quantified 2 244

Detected and quantified please fill in Q6a-6h 80 976

Found absent 0 0

No Answer 0 0

T2 7 Select the option applicable to your analysis for 68416 soybean in T2

Answers Ratio

Not tested 15 183

Detected but not quantified 1 122

Detected and quantified please fill in Q7a-7h 0 0

Found absent 66 805

No Answer 0 0

T2 8 Select the option applicable to your analysis for MON89788 soybean in T2

Answers Ratio

Not tested 5 61

Detected but not quantified 6 732

Detected and quantified please fill in Q8a-8h 3 366

Found absent 68 829

No Answer 0 0

T2 9 If applicable why did you not test or quantify all GM events in T2

Answers Ratio

a) Not applicable all GM events listed were tested and all those detected were quantified 61 744

b) The event-specific detection method is not validated in our laboratory 6 732

c) Reference material primers probes or other reagents were not available (in time) 10 122

d) The result obtained was below the LODLOQ 5 61

e) Practical constraints (instrument broken no personnel etc) 0 0

f) Other reason 2 244

No Answer 0 0

T2 6a 40-3-2 Which quantification approach was used

Answers Ratio

Standard curve method (2 calibration curves) 71 866

Delta Cq method (one calibration curve) 9 110

Digital PCR (no calibration curve) 2 244

EURL GMFF CT 0217 report

3646

No Answer 2 244

T2 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 74 902

CRM from IRMM certified in GM copy number ratio (plasmid CRM) 0 0

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 2 244

T2 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T2

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 63 768

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 6 732

Other please specify below 2 244

No Answer 5 61

Specify the reference target(s) used (if different from above)

Lectin - 74bp

lec 74bp

Lectine

SOJA LEKTIN 80 bp Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

Terry C F Harris N Event-specific detection of Roundup Ready soya using two different real time PCR detection chemistries Eur Food Res Technol (2001) 213425-431

T2 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

61 744

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

14 171

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 2 244

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T2 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 7 854

100 ng 21 256

EURL GMFF CT 0217 report

3746

50 ng 17 207

25 ng 7 854

15 ng 2 244

lt10 ng 0 0

No Answer 2 244

T2 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 5 002 6 003 1 004 7 005 7 006 2 007 1 008 2 009 8 01 39 026 1 028 1

T2 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 21 256

Determined during the in-house validation of the method 43 524

Taken from the EURL GMFF validation report 19 232

By another approach please explain below 3 366

No Answer 2 244

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample

Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit

T2 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 30 366

Known uncertainty of the standard method 8 976

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 7 854

No Answer 2 244

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

We used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

30 of the quantification result

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

Event 68416 was not tested due to non-availability of Positive control and reference material

DNA extraction of T2 has conducted using DNA extraction kits SureFood Prep Basic etc (see T1)

The NucleoSpin Food + NucleoSpin gDNA Clean-up was also used for DNA extraction from T1 and T2 samples The results for 40-3-2 were 25 - 30 lower than CTAB results T1 050 +- 018 (ww) T2 058 +- 020 (ww)

AOCS 0906-B as calibrant for MON89788 in T2 also

EURL GMFF CT 0217 report

3846

Annex 3 Qualitative performance of the participants

Table A31 Performance of the participants for the qualitative identification of GM events in CT-0217

The correct answer (yes) is that the GM event has been detected in the test item

Lab 40-3-2 detected in T1 40-3-2 detected in T2 Lab 40-3-2 detected in T1 40-3-2 detected in T2

L01 yes yes L45 yes yes

L02 yes yes L46 yes yes

L03 yes yes L47 yes yes

L04 yes yes L48 yes yes

L05 yes yes L49 yes yes

L06 yes yes L50 yes yes

L07 yes yes L51 yes yes

L08 yes yes L52 yes yes

L09 yes yes L53 yes yes

L10 yes Yes L54 yes yes

L11 yes yes L55 yes yes

L12 yes yes L56 yes yes

L13 yes yes L57 yes yes

L14 yes yes L58 yes yes

L15 yes yes L59 yes yes

L16 yes yes L60 yes yes

L17 yes yes L61 yes yes

L18 yes yes L62 yes yes

L19 yes yes L63 yes yes

L20 yes yes L66 yes yes

L22 yes yes L68 yes yes

L23 yes yes L69 yes yes

L24 yes yes L70 yes yes

L25 yes yes L71 yes yes

L26 yes yes L72 yes yes

L27 yes yes L73 T1 was not analysed yes

L29 yes yes L74 yes yes

L30 yes yes L75 yes yes

L31 yes yes L76 yes yes

L32 yes yes L77 yes yes

L33 yes yes L78 yes yes

L34 yes yes L79 yes yes

L35 yes yes L80 yes yes

L36 yes yes L81 yes yes

L37 yes yes L82 yes yes

L38 yes yes L83 yes yes

L39 yes yes L84 yes yes

L40 yes yes L85 yes yes

L41 yes yes L86 yes yes

L42 yes yes L87 yes yes

L43 T1 was not analysed yes L88 yes yes

L44 yes yes L89 yes yes

Although the questionnaire was not returned by the participant the presence of the event was inferred from the quantitative result reported

L10 did not fill in the questionnaire but reported a larger than 01 result for T1 and T2

EURL GMFF CT 0217 report

3946

Annex 4 Participants quantitative performance

The z and ζ scores of all laboratories are reported in Tables A41 and A42 and in Figures

A41 and A42 for 40-3-2 soybean in T1 and T2 respectively For consistency the

reported results are shown with two decimals the performance scores with one decimal

The z scores are displayed in green (satisfactory) orange (questionable) or red cells

(unsatisfactory)

Table A41 Quantitative results (in mm ) and performance scores of participants of CT-0217 for 40-3-2 soybean in chicken feed (T1)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 064 008 200 -019 004 003 -10 -28

L02 NRL882 057 002 200 -024 001 001 -15 -67

L03 Non-NRL 073 1520 200 -014 760 452 -04 00

L04 Non-NRL 075

-012 000 000 -03 -13

L05 NRL120 073 007 257 -014 003 002 -04 -15

L06 NRL882 075 021 200 -012 011 006 -03 -04

L07 NRL882 065 013 200 -019 007 004 -09 -19

L08 NRL120 062 023 200 -021 012 008 -11 -13

L09 NRL882 081 027 200 -009 014 007 01 01

L12 Non-NRL 100 051 200 000 026 011 10 09

L13 NRL882 094 019 173 -003 011 005 07 13

L14 Non-NRL 063

-020 000 000 -10 -50

L15 NRL882 110 011 200 004 005 002 14 46

L16 Non-NRL 117

007 000 000 17 80

L17 NRL882 072 010 200 -014 005 003 -05 -12

L18 Non-NRL 040

-040 000 000 -30 -145

L19 Non-NRL 115 063 173 006 036 014 16 11

L20 NRL120 108 032 200 003 016 006 13 19

L22 Non-NRL 051 030 223 -029 013 011 -20 -17

L23 Non-NRL 047 009 200 -033 005 004 -23 -50

L24 NRL882 072 019 200 -014 010 006 -05 -07

L25 Non-NRL 120

008 000 000 18 85

L26 NRL882 023 009 200 -064 005 008 -54 -62

L27 Non-NRL 098 015 200 -001 008 003 09 23

L29 NRL120 090 008 173 -005 005 002 05 17

L30 NRL882 045 013 200 -035 007 006 -25 -38

L31 NRL882 080 038 223 -010 017 009 00 00

L32 Non-NRL 076 039 200 -012 020 011 -02 -02

L33 NRL-120 057 014 200 -024 007 005 -15 -26

L34 NRL120 061 016 200 -021 008 006 -12 -19

L35 NRL882 097 027 200 -001 014 006 08 13

L36 Non-NRL 090 030 200 -005 015 007 05 07

L37 NRL120 065 026 200 -019 013 009 -09 -10

L38 NRL120 093 014 237 -003 006 003 07 19

L39 Non-NRL 112

005 000 000 15 71

L40 NRL882 053 008 200 -028 004 003 -18 -46

L41 NRL120 080 003 200 -010 002 001 00 00

L42 NRL882 104 031 200 002 016 006 11 17

L44 NRL120 084 007 257 -008 003 001 02 09

L45 NRL882 065 018 200 -019 009 006 -09 -14

L46 NRL882 081 018 200 -009 009 005 01 01

L47 NRL882 141 027 173 015 016 005 25 47

L48 NRL882 059 028 200 -023 014 010 -13 -13

L49 Non-NRL 089 031 200 -005 016 008 05 06

L50 NRL120 073 005 200 -014 003 001 -04 -15

L51 Non-NRL 077 049 200 -011 025 014 -02 -01

L52 Non-NRL 091 012 200 -004 006 003 06 16

L53 Non-NRL 043 081 200 -037 041 041 -27 -07

L54 NRL882 074 035 200 -013 018 010 -03 -03

L55 Non-NRL 076 019 173 -012 011 006 -02 -03

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

3046

Annex 2 Questionnaire data

The results received from 82 laboratories were exported from the EUSurvey

Questionnaire on CT 0217 analysis and are tabulated below Multiple answers were

allowed for all questions except for the questions on the calibrant used The results of the

open questions were manually analysed and reported Answers to the questions on GM

events that were not to be quantified in the test items are not shown

Select the group to which your organisation belongs Note 882 and 120 refer to EU Regulations 8822004 and 1202014

resp select NRL120 if your organisation is ONLY listed under Regulation 1202014 select non-NRL if your organisation is

not an NRL under either EU Regulation

Answers Ratio

NRL882 33 402

NRL120 20 244

Non-NRL 29 354

No Answer 0 0

T1 Please select the option that applies and proceed with the questionnaire

Answers Ratio

T1 was not analysed go to Q1 2 244

T1 was analysed go to Q2 80 976

No Answer 0 0

T1 1 Why did you not analyse test item 1

Answers Ratio

a) The sample matrix is out of the scope of our laboratory 1 122

b) The methods are not validated in our laboratory 0 0

c) We could not obtain sufficient good quality DNA suitable for further analysis 0 0

d) Reference material primers probes or other reagents were not available (in time) 0 0

e) We tried but our analysis failed 0 0

f) Other practical constraints (instrument broken no personnel etc) 1 122

g) Other reason 0 0

No Answer 80 976

T1 2 Select the DNA extraction method used for T1

Answers Ratio

CTAB 41 50

NucleoSpin Food 10 122

NucleoSpin Plant 3 366

GeneSpin 4 488

Promega Wizard 3 366

DNeasy Plant 3 366

DNeasy Mericon Food 5 61

Biotecon Foodproof 5 61

SDS 4 488

Fast ID Genomic DNA 2 244

Maxwell 16 Plant DNA 0 0

Maxwell 16 Food Feed Seed 5 61

Generon Ion Force 1 122

Other 2 244

No Answer 2 244

T1 3 Select any additional DNA purification method used for T1

Answers Ratio

No additional clean-up 51 622

Additional ethanol precipitation 9 110

Eurofins DNAExtractor cleaning column 3 366

Promega Wizard DNA clean-up resin 6 732

Qiagen QIAQuick 4 488

Qiagen Genomic-Tip 20G 1 122

Other method (no need to specify) 7 854

No Answer 2 244

EURL GMFF CT 0217 report

3146

T1 4 Indicate the number of replicate DNA extractions used to obtain the results

Answers Ratio

1 0 0

2 52 634

3 9 110

4 10 122

5 2 244

6 6 732

gt6 1 122

No Answer 2 244

T1 5 Select the approach(es) used to show absence of PCR inhibition

Answers Ratio None (no inhibition was suspected based on experience) 6 732

We check that the optical density ratios (OD260280 260230) are acceptable 40 488

We verify that the amplification curves look normal 24 293

We run two dilutions and verify if the delta Cq or GM are as expected 31 378

We run three or four dilutions and verify if the delta Cq or GM are as expected 13 158

We perform a PCR inhibition run with a reference gene before analysis 3 or 4 dilutions linear regression extrapolation of Cq for undiluted extract compare this to the measured Cq

16 195

We add an internal positive control to the reactions and check the Cq 15 183

Other 1 122

No Answer 2 244

T1 6 Select the option applicable to your analysis for 40-3-2 soybean in T1

Answers Ratio

Not tested 0 0

Detected but not quantified 3 366

Detected and quantified please fill in Q6a-6h 77 939

Found absent 0 0

No Answer 2 244

T1 7 Select the option applicable to your analysis for 68416 soybean in T1

Answers Ratio

Not tested 14 171

Detected but not quantified 2 244

Detected and quantified please fill in Q7a-7h 0 0

Found absent 64 781

No Answer 2 244

T1 8 Select the option applicable to your analysis for MON89788 soybean in T1

Answers Ratio

Not tested 6 732

Detected but not quantified 33 402

Detected and quantified please fill in Q8a-8h 15 183

Found absent 26 317

No Answer 2 244

T1 9 If applicable why did you not test or quantify all GM events in T1

Answers Ratio

a) Not applicable all GM events listed were tested and all those detected were quantified 45 549

b) The event-specific detection method is not validated in our laboratory 7 854

c) Reference material primers probes or other reagents were not available (in time) 9 110

d) The result obtained was below the LODLOQ 18 220

e) Practical constraints (instrument broken no personnel etc) 0 0

f) Other reason 9 110

No Answer 2 244

T1 6a Soybean 40-3-2 Which quantification approach was used

Answers Ratio

Standard curve method (2 calibration curves) 68 829

Delta Cq method (one calibration curve) 9 110

Digital PCR (no calibration curve) 2 244

EURL GMFF CT 0217 report

3246

No Answer 5 61

T1 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 71 866

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 5 61

T1 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T1

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 65 793

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 5 61

Other please specify below 1 122

No Answer 5 61

Specify the reference target(s) used (if different from above)

SOJA LEKTIN 80 bp - Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

T1 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

59 720

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

13 159

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 5 61

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T1 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 9 110

100 ng 18 220

50 ng 14 171

25 ng 6 732

15 ng 0 0

lt10 ng 0 0

No Answer 5 61

EURL GMFF CT 0217 report

3346

T1 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 3 002 4 003 1 004 6 005 7 006 2 008 4 009 7 01 37 019 1 02 1 03 1 035 1 04 1 13 1

T1 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 20 244

Determined during the in-house validation of the method 42 512

Taken from the EURL GMFF validation report 17 207

By another approach please explain below 2 244

No Answer 5 61

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

T1 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2 soybean

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 27 329

Known uncertainty of the standard method 9 110

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 6 732

No Answer 5 61

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx

Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

we used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

MON89788 soybean was detected but not quantified because results for this event were lt LOQ

We saw in the qPCR for MON89788 Ct-values at 37 - 39 but lt LOD (lt 002 )

Additional tests with one extract performed with Nucleospin Food Kit and with one extract performed with CTAB+Nucleospin Food Kit showed no significant differences in event detection and GTS 40-3-2 quantification (examined with digital PCR)

T19 not all detected GM were asked to be quantified

For T1 trace amounts of MON89788 were detected either in qualitative screening as well as in quantification The givencalculated GM (mm) of 002 is well below the known and validated LOQ of the method which is 01 We detected a low level of MON89788 but as this quantified at 0006 lower than the threshold for reporting we took this as a negative result

According course material of JRC GMO Quantification Proper calibration and Estimation of Measurement Uncertainty

MU is based on intermediate precision data

EURL GMFF CT 0217 report

3446

The sample was strongly inhibited MON89788 was detected under LOQ

In T1 the measured GM for MON89788 of 001 is below the determined LOQ (02) therefore no MU was reported

MON89788 was detected in traces at the LOD

MON89788 SOY result obtained below the LOQ but copy numbers of MON89788 was above LOD

T1 was inhibited A 10x dilution of the DNA was needed to get a reasonable result

DNA extraction of T1 has conducted using DNA extraction kits SureFood Prep Basic (S1052) and SureFood Prep Advanced (S1053) R-Biopharm AG Identification of Soya GM-lines was conducted using diagnostic kits SureFood GMO ID Roundup Ready Soya (S2030) and SureFood GMO ID RR2Y (S2034) Quantification kits were SureFood GMO Quant Roundup Ready Soya (S2014) SureFood GMO Quant 35S soya (S2028) and SureFood GMO Quant RR2Y Soya (S2029) R-Biopharm AG MON89788 soybean was detected in sample T1 but only in some cases - 2 PCR replicates were always positive out of 4 after several repetitions

Remark (precision) in relation to T18 and T19 MON89788 soybean was detected in the sample but not quantified because of lt LOQ

T2 Please select the option that applies and proceed with the questionnaire

Answers Ratio

T2 was not analysed go to Q1 0 0

T2 was analysed go to Q2 82 100

No Answer 0 0

T2 1 Why did you not analyse test item 2

Answers Ratio

a) The sample matrix is out of the scope of our laboratory 0 0

b) The methods are not validated in our laboratory 0 0

c) We could not obtain sufficient good quality DNA suitable for further analysis 0 0

d) Reference material primers probes or other reagents were not available (in time) 0 0

e) We tried but our analysis failed 0 0

f) Other practical constraints (instrument broken no personnel etc) 0 0

g) Other reason 0 0

No Answer 82 100

T2 2 Select the DNA extraction method used for T2

Answers Ratio

CTAB 40 488

NucleoSpin Food 11 134

NucleoSpin Plant 3 366

GeneSpin 4 488

Promega Wizard 3 366

DNeasy Plant 3 366

DNeasy Mericon Food 4 488

Biotecon Foodproof 5 61

SDS 4 488

Fast ID Genomic DNA 3 366

Maxwell 16 Plant DNA 0 0

Maxwell 16 Food Feed Seed 4 488

Generon Ion Force 1 122

Other 2 244

No Answer 0 0

T2 3 Select any additional DNA purification method used for T2

Answers Ratio

No additional clean-up 52 634

Additional ethanol precipitation 9 110

Eurofins DNAExtractor cleaning column 3 366

Promega Wizard DNA clean-up resin 6 732

Qiagen QIAQuick 5 61

Qiagen Genomic-Tip 20G 1 122

Other method (no need to specify) 8 976

No Answer 0 0

EURL GMFF CT 0217 report

3546

T2 4 Indicate the number of replicate DNA extractions used to obtain the results

Answers Ratio

1 0 0

2 56 683

3 9 110

4 10 122

5 2 244

6 4 488

gt6 1 122

No Answer 0 0

T2 5 Select the approach(es) used to show absence of PCR inhibition

Answers Ratio

None (no inhibition was suspected based on experience) 7 854

We run two dilutions and verify if the delta Cq or GM are as expected 32 390

We run three or four dilutions and verify if the delta Cq or GM are as expected 12 146

We perform a PCR inhibition run with a reference gene before analysis 3 or 4 dilutions linear regression extrapolation of Cq of undiluted extract compare this to the measured Cq

15 183

We add an internal positive control to the reactions and check the Cq 16 195

We verify that the amplification curves look normal 25 305

We check that the optical density ratios (OD260280 260230) are acceptable 36 439

Other 1 122

No Answer 0 0

T2 6 Select the option applicable to your analysis for 40-3-2 soybean in T2

Answers Ratio

Not tested 0 0

Detected but not quantified 2 244

Detected and quantified please fill in Q6a-6h 80 976

Found absent 0 0

No Answer 0 0

T2 7 Select the option applicable to your analysis for 68416 soybean in T2

Answers Ratio

Not tested 15 183

Detected but not quantified 1 122

Detected and quantified please fill in Q7a-7h 0 0

Found absent 66 805

No Answer 0 0

T2 8 Select the option applicable to your analysis for MON89788 soybean in T2

Answers Ratio

Not tested 5 61

Detected but not quantified 6 732

Detected and quantified please fill in Q8a-8h 3 366

Found absent 68 829

No Answer 0 0

T2 9 If applicable why did you not test or quantify all GM events in T2

Answers Ratio

a) Not applicable all GM events listed were tested and all those detected were quantified 61 744

b) The event-specific detection method is not validated in our laboratory 6 732

c) Reference material primers probes or other reagents were not available (in time) 10 122

d) The result obtained was below the LODLOQ 5 61

e) Practical constraints (instrument broken no personnel etc) 0 0

f) Other reason 2 244

No Answer 0 0

T2 6a 40-3-2 Which quantification approach was used

Answers Ratio

Standard curve method (2 calibration curves) 71 866

Delta Cq method (one calibration curve) 9 110

Digital PCR (no calibration curve) 2 244

EURL GMFF CT 0217 report

3646

No Answer 2 244

T2 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 74 902

CRM from IRMM certified in GM copy number ratio (plasmid CRM) 0 0

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 2 244

T2 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T2

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 63 768

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 6 732

Other please specify below 2 244

No Answer 5 61

Specify the reference target(s) used (if different from above)

Lectin - 74bp

lec 74bp

Lectine

SOJA LEKTIN 80 bp Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

Terry C F Harris N Event-specific detection of Roundup Ready soya using two different real time PCR detection chemistries Eur Food Res Technol (2001) 213425-431

T2 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

61 744

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

14 171

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 2 244

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T2 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 7 854

100 ng 21 256

EURL GMFF CT 0217 report

3746

50 ng 17 207

25 ng 7 854

15 ng 2 244

lt10 ng 0 0

No Answer 2 244

T2 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 5 002 6 003 1 004 7 005 7 006 2 007 1 008 2 009 8 01 39 026 1 028 1

T2 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 21 256

Determined during the in-house validation of the method 43 524

Taken from the EURL GMFF validation report 19 232

By another approach please explain below 3 366

No Answer 2 244

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample

Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit

T2 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 30 366

Known uncertainty of the standard method 8 976

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 7 854

No Answer 2 244

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

We used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

30 of the quantification result

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

Event 68416 was not tested due to non-availability of Positive control and reference material

DNA extraction of T2 has conducted using DNA extraction kits SureFood Prep Basic etc (see T1)

The NucleoSpin Food + NucleoSpin gDNA Clean-up was also used for DNA extraction from T1 and T2 samples The results for 40-3-2 were 25 - 30 lower than CTAB results T1 050 +- 018 (ww) T2 058 +- 020 (ww)

AOCS 0906-B as calibrant for MON89788 in T2 also

EURL GMFF CT 0217 report

3846

Annex 3 Qualitative performance of the participants

Table A31 Performance of the participants for the qualitative identification of GM events in CT-0217

The correct answer (yes) is that the GM event has been detected in the test item

Lab 40-3-2 detected in T1 40-3-2 detected in T2 Lab 40-3-2 detected in T1 40-3-2 detected in T2

L01 yes yes L45 yes yes

L02 yes yes L46 yes yes

L03 yes yes L47 yes yes

L04 yes yes L48 yes yes

L05 yes yes L49 yes yes

L06 yes yes L50 yes yes

L07 yes yes L51 yes yes

L08 yes yes L52 yes yes

L09 yes yes L53 yes yes

L10 yes Yes L54 yes yes

L11 yes yes L55 yes yes

L12 yes yes L56 yes yes

L13 yes yes L57 yes yes

L14 yes yes L58 yes yes

L15 yes yes L59 yes yes

L16 yes yes L60 yes yes

L17 yes yes L61 yes yes

L18 yes yes L62 yes yes

L19 yes yes L63 yes yes

L20 yes yes L66 yes yes

L22 yes yes L68 yes yes

L23 yes yes L69 yes yes

L24 yes yes L70 yes yes

L25 yes yes L71 yes yes

L26 yes yes L72 yes yes

L27 yes yes L73 T1 was not analysed yes

L29 yes yes L74 yes yes

L30 yes yes L75 yes yes

L31 yes yes L76 yes yes

L32 yes yes L77 yes yes

L33 yes yes L78 yes yes

L34 yes yes L79 yes yes

L35 yes yes L80 yes yes

L36 yes yes L81 yes yes

L37 yes yes L82 yes yes

L38 yes yes L83 yes yes

L39 yes yes L84 yes yes

L40 yes yes L85 yes yes

L41 yes yes L86 yes yes

L42 yes yes L87 yes yes

L43 T1 was not analysed yes L88 yes yes

L44 yes yes L89 yes yes

Although the questionnaire was not returned by the participant the presence of the event was inferred from the quantitative result reported

L10 did not fill in the questionnaire but reported a larger than 01 result for T1 and T2

EURL GMFF CT 0217 report

3946

Annex 4 Participants quantitative performance

The z and ζ scores of all laboratories are reported in Tables A41 and A42 and in Figures

A41 and A42 for 40-3-2 soybean in T1 and T2 respectively For consistency the

reported results are shown with two decimals the performance scores with one decimal

The z scores are displayed in green (satisfactory) orange (questionable) or red cells

(unsatisfactory)

Table A41 Quantitative results (in mm ) and performance scores of participants of CT-0217 for 40-3-2 soybean in chicken feed (T1)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 064 008 200 -019 004 003 -10 -28

L02 NRL882 057 002 200 -024 001 001 -15 -67

L03 Non-NRL 073 1520 200 -014 760 452 -04 00

L04 Non-NRL 075

-012 000 000 -03 -13

L05 NRL120 073 007 257 -014 003 002 -04 -15

L06 NRL882 075 021 200 -012 011 006 -03 -04

L07 NRL882 065 013 200 -019 007 004 -09 -19

L08 NRL120 062 023 200 -021 012 008 -11 -13

L09 NRL882 081 027 200 -009 014 007 01 01

L12 Non-NRL 100 051 200 000 026 011 10 09

L13 NRL882 094 019 173 -003 011 005 07 13

L14 Non-NRL 063

-020 000 000 -10 -50

L15 NRL882 110 011 200 004 005 002 14 46

L16 Non-NRL 117

007 000 000 17 80

L17 NRL882 072 010 200 -014 005 003 -05 -12

L18 Non-NRL 040

-040 000 000 -30 -145

L19 Non-NRL 115 063 173 006 036 014 16 11

L20 NRL120 108 032 200 003 016 006 13 19

L22 Non-NRL 051 030 223 -029 013 011 -20 -17

L23 Non-NRL 047 009 200 -033 005 004 -23 -50

L24 NRL882 072 019 200 -014 010 006 -05 -07

L25 Non-NRL 120

008 000 000 18 85

L26 NRL882 023 009 200 -064 005 008 -54 -62

L27 Non-NRL 098 015 200 -001 008 003 09 23

L29 NRL120 090 008 173 -005 005 002 05 17

L30 NRL882 045 013 200 -035 007 006 -25 -38

L31 NRL882 080 038 223 -010 017 009 00 00

L32 Non-NRL 076 039 200 -012 020 011 -02 -02

L33 NRL-120 057 014 200 -024 007 005 -15 -26

L34 NRL120 061 016 200 -021 008 006 -12 -19

L35 NRL882 097 027 200 -001 014 006 08 13

L36 Non-NRL 090 030 200 -005 015 007 05 07

L37 NRL120 065 026 200 -019 013 009 -09 -10

L38 NRL120 093 014 237 -003 006 003 07 19

L39 Non-NRL 112

005 000 000 15 71

L40 NRL882 053 008 200 -028 004 003 -18 -46

L41 NRL120 080 003 200 -010 002 001 00 00

L42 NRL882 104 031 200 002 016 006 11 17

L44 NRL120 084 007 257 -008 003 001 02 09

L45 NRL882 065 018 200 -019 009 006 -09 -14

L46 NRL882 081 018 200 -009 009 005 01 01

L47 NRL882 141 027 173 015 016 005 25 47

L48 NRL882 059 028 200 -023 014 010 -13 -13

L49 Non-NRL 089 031 200 -005 016 008 05 06

L50 NRL120 073 005 200 -014 003 001 -04 -15

L51 Non-NRL 077 049 200 -011 025 014 -02 -01

L52 Non-NRL 091 012 200 -004 006 003 06 16

L53 Non-NRL 043 081 200 -037 041 041 -27 -07

L54 NRL882 074 035 200 -013 018 010 -03 -03

L55 Non-NRL 076 019 173 -012 011 006 -02 -03

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

3146

T1 4 Indicate the number of replicate DNA extractions used to obtain the results

Answers Ratio

1 0 0

2 52 634

3 9 110

4 10 122

5 2 244

6 6 732

gt6 1 122

No Answer 2 244

T1 5 Select the approach(es) used to show absence of PCR inhibition

Answers Ratio None (no inhibition was suspected based on experience) 6 732

We check that the optical density ratios (OD260280 260230) are acceptable 40 488

We verify that the amplification curves look normal 24 293

We run two dilutions and verify if the delta Cq or GM are as expected 31 378

We run three or four dilutions and verify if the delta Cq or GM are as expected 13 158

We perform a PCR inhibition run with a reference gene before analysis 3 or 4 dilutions linear regression extrapolation of Cq for undiluted extract compare this to the measured Cq

16 195

We add an internal positive control to the reactions and check the Cq 15 183

Other 1 122

No Answer 2 244

T1 6 Select the option applicable to your analysis for 40-3-2 soybean in T1

Answers Ratio

Not tested 0 0

Detected but not quantified 3 366

Detected and quantified please fill in Q6a-6h 77 939

Found absent 0 0

No Answer 2 244

T1 7 Select the option applicable to your analysis for 68416 soybean in T1

Answers Ratio

Not tested 14 171

Detected but not quantified 2 244

Detected and quantified please fill in Q7a-7h 0 0

Found absent 64 781

No Answer 2 244

T1 8 Select the option applicable to your analysis for MON89788 soybean in T1

Answers Ratio

Not tested 6 732

Detected but not quantified 33 402

Detected and quantified please fill in Q8a-8h 15 183

Found absent 26 317

No Answer 2 244

T1 9 If applicable why did you not test or quantify all GM events in T1

Answers Ratio

a) Not applicable all GM events listed were tested and all those detected were quantified 45 549

b) The event-specific detection method is not validated in our laboratory 7 854

c) Reference material primers probes or other reagents were not available (in time) 9 110

d) The result obtained was below the LODLOQ 18 220

e) Practical constraints (instrument broken no personnel etc) 0 0

f) Other reason 9 110

No Answer 2 244

T1 6a Soybean 40-3-2 Which quantification approach was used

Answers Ratio

Standard curve method (2 calibration curves) 68 829

Delta Cq method (one calibration curve) 9 110

Digital PCR (no calibration curve) 2 244

EURL GMFF CT 0217 report

3246

No Answer 5 61

T1 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 71 866

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 5 61

T1 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T1

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 65 793

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 5 61

Other please specify below 1 122

No Answer 5 61

Specify the reference target(s) used (if different from above)

SOJA LEKTIN 80 bp - Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

T1 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

59 720

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

13 159

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 5 61

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T1 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 9 110

100 ng 18 220

50 ng 14 171

25 ng 6 732

15 ng 0 0

lt10 ng 0 0

No Answer 5 61

EURL GMFF CT 0217 report

3346

T1 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 3 002 4 003 1 004 6 005 7 006 2 008 4 009 7 01 37 019 1 02 1 03 1 035 1 04 1 13 1

T1 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 20 244

Determined during the in-house validation of the method 42 512

Taken from the EURL GMFF validation report 17 207

By another approach please explain below 2 244

No Answer 5 61

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

T1 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2 soybean

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 27 329

Known uncertainty of the standard method 9 110

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 6 732

No Answer 5 61

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx

Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

we used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

MON89788 soybean was detected but not quantified because results for this event were lt LOQ

We saw in the qPCR for MON89788 Ct-values at 37 - 39 but lt LOD (lt 002 )

Additional tests with one extract performed with Nucleospin Food Kit and with one extract performed with CTAB+Nucleospin Food Kit showed no significant differences in event detection and GTS 40-3-2 quantification (examined with digital PCR)

T19 not all detected GM were asked to be quantified

For T1 trace amounts of MON89788 were detected either in qualitative screening as well as in quantification The givencalculated GM (mm) of 002 is well below the known and validated LOQ of the method which is 01 We detected a low level of MON89788 but as this quantified at 0006 lower than the threshold for reporting we took this as a negative result

According course material of JRC GMO Quantification Proper calibration and Estimation of Measurement Uncertainty

MU is based on intermediate precision data

EURL GMFF CT 0217 report

3446

The sample was strongly inhibited MON89788 was detected under LOQ

In T1 the measured GM for MON89788 of 001 is below the determined LOQ (02) therefore no MU was reported

MON89788 was detected in traces at the LOD

MON89788 SOY result obtained below the LOQ but copy numbers of MON89788 was above LOD

T1 was inhibited A 10x dilution of the DNA was needed to get a reasonable result

DNA extraction of T1 has conducted using DNA extraction kits SureFood Prep Basic (S1052) and SureFood Prep Advanced (S1053) R-Biopharm AG Identification of Soya GM-lines was conducted using diagnostic kits SureFood GMO ID Roundup Ready Soya (S2030) and SureFood GMO ID RR2Y (S2034) Quantification kits were SureFood GMO Quant Roundup Ready Soya (S2014) SureFood GMO Quant 35S soya (S2028) and SureFood GMO Quant RR2Y Soya (S2029) R-Biopharm AG MON89788 soybean was detected in sample T1 but only in some cases - 2 PCR replicates were always positive out of 4 after several repetitions

Remark (precision) in relation to T18 and T19 MON89788 soybean was detected in the sample but not quantified because of lt LOQ

T2 Please select the option that applies and proceed with the questionnaire

Answers Ratio

T2 was not analysed go to Q1 0 0

T2 was analysed go to Q2 82 100

No Answer 0 0

T2 1 Why did you not analyse test item 2

Answers Ratio

a) The sample matrix is out of the scope of our laboratory 0 0

b) The methods are not validated in our laboratory 0 0

c) We could not obtain sufficient good quality DNA suitable for further analysis 0 0

d) Reference material primers probes or other reagents were not available (in time) 0 0

e) We tried but our analysis failed 0 0

f) Other practical constraints (instrument broken no personnel etc) 0 0

g) Other reason 0 0

No Answer 82 100

T2 2 Select the DNA extraction method used for T2

Answers Ratio

CTAB 40 488

NucleoSpin Food 11 134

NucleoSpin Plant 3 366

GeneSpin 4 488

Promega Wizard 3 366

DNeasy Plant 3 366

DNeasy Mericon Food 4 488

Biotecon Foodproof 5 61

SDS 4 488

Fast ID Genomic DNA 3 366

Maxwell 16 Plant DNA 0 0

Maxwell 16 Food Feed Seed 4 488

Generon Ion Force 1 122

Other 2 244

No Answer 0 0

T2 3 Select any additional DNA purification method used for T2

Answers Ratio

No additional clean-up 52 634

Additional ethanol precipitation 9 110

Eurofins DNAExtractor cleaning column 3 366

Promega Wizard DNA clean-up resin 6 732

Qiagen QIAQuick 5 61

Qiagen Genomic-Tip 20G 1 122

Other method (no need to specify) 8 976

No Answer 0 0

EURL GMFF CT 0217 report

3546

T2 4 Indicate the number of replicate DNA extractions used to obtain the results

Answers Ratio

1 0 0

2 56 683

3 9 110

4 10 122

5 2 244

6 4 488

gt6 1 122

No Answer 0 0

T2 5 Select the approach(es) used to show absence of PCR inhibition

Answers Ratio

None (no inhibition was suspected based on experience) 7 854

We run two dilutions and verify if the delta Cq or GM are as expected 32 390

We run three or four dilutions and verify if the delta Cq or GM are as expected 12 146

We perform a PCR inhibition run with a reference gene before analysis 3 or 4 dilutions linear regression extrapolation of Cq of undiluted extract compare this to the measured Cq

15 183

We add an internal positive control to the reactions and check the Cq 16 195

We verify that the amplification curves look normal 25 305

We check that the optical density ratios (OD260280 260230) are acceptable 36 439

Other 1 122

No Answer 0 0

T2 6 Select the option applicable to your analysis for 40-3-2 soybean in T2

Answers Ratio

Not tested 0 0

Detected but not quantified 2 244

Detected and quantified please fill in Q6a-6h 80 976

Found absent 0 0

No Answer 0 0

T2 7 Select the option applicable to your analysis for 68416 soybean in T2

Answers Ratio

Not tested 15 183

Detected but not quantified 1 122

Detected and quantified please fill in Q7a-7h 0 0

Found absent 66 805

No Answer 0 0

T2 8 Select the option applicable to your analysis for MON89788 soybean in T2

Answers Ratio

Not tested 5 61

Detected but not quantified 6 732

Detected and quantified please fill in Q8a-8h 3 366

Found absent 68 829

No Answer 0 0

T2 9 If applicable why did you not test or quantify all GM events in T2

Answers Ratio

a) Not applicable all GM events listed were tested and all those detected were quantified 61 744

b) The event-specific detection method is not validated in our laboratory 6 732

c) Reference material primers probes or other reagents were not available (in time) 10 122

d) The result obtained was below the LODLOQ 5 61

e) Practical constraints (instrument broken no personnel etc) 0 0

f) Other reason 2 244

No Answer 0 0

T2 6a 40-3-2 Which quantification approach was used

Answers Ratio

Standard curve method (2 calibration curves) 71 866

Delta Cq method (one calibration curve) 9 110

Digital PCR (no calibration curve) 2 244

EURL GMFF CT 0217 report

3646

No Answer 2 244

T2 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 74 902

CRM from IRMM certified in GM copy number ratio (plasmid CRM) 0 0

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 2 244

T2 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T2

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 63 768

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 6 732

Other please specify below 2 244

No Answer 5 61

Specify the reference target(s) used (if different from above)

Lectin - 74bp

lec 74bp

Lectine

SOJA LEKTIN 80 bp Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

Terry C F Harris N Event-specific detection of Roundup Ready soya using two different real time PCR detection chemistries Eur Food Res Technol (2001) 213425-431

T2 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

61 744

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

14 171

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 2 244

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T2 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 7 854

100 ng 21 256

EURL GMFF CT 0217 report

3746

50 ng 17 207

25 ng 7 854

15 ng 2 244

lt10 ng 0 0

No Answer 2 244

T2 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 5 002 6 003 1 004 7 005 7 006 2 007 1 008 2 009 8 01 39 026 1 028 1

T2 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 21 256

Determined during the in-house validation of the method 43 524

Taken from the EURL GMFF validation report 19 232

By another approach please explain below 3 366

No Answer 2 244

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample

Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit

T2 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 30 366

Known uncertainty of the standard method 8 976

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 7 854

No Answer 2 244

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

We used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

30 of the quantification result

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

Event 68416 was not tested due to non-availability of Positive control and reference material

DNA extraction of T2 has conducted using DNA extraction kits SureFood Prep Basic etc (see T1)

The NucleoSpin Food + NucleoSpin gDNA Clean-up was also used for DNA extraction from T1 and T2 samples The results for 40-3-2 were 25 - 30 lower than CTAB results T1 050 +- 018 (ww) T2 058 +- 020 (ww)

AOCS 0906-B as calibrant for MON89788 in T2 also

EURL GMFF CT 0217 report

3846

Annex 3 Qualitative performance of the participants

Table A31 Performance of the participants for the qualitative identification of GM events in CT-0217

The correct answer (yes) is that the GM event has been detected in the test item

Lab 40-3-2 detected in T1 40-3-2 detected in T2 Lab 40-3-2 detected in T1 40-3-2 detected in T2

L01 yes yes L45 yes yes

L02 yes yes L46 yes yes

L03 yes yes L47 yes yes

L04 yes yes L48 yes yes

L05 yes yes L49 yes yes

L06 yes yes L50 yes yes

L07 yes yes L51 yes yes

L08 yes yes L52 yes yes

L09 yes yes L53 yes yes

L10 yes Yes L54 yes yes

L11 yes yes L55 yes yes

L12 yes yes L56 yes yes

L13 yes yes L57 yes yes

L14 yes yes L58 yes yes

L15 yes yes L59 yes yes

L16 yes yes L60 yes yes

L17 yes yes L61 yes yes

L18 yes yes L62 yes yes

L19 yes yes L63 yes yes

L20 yes yes L66 yes yes

L22 yes yes L68 yes yes

L23 yes yes L69 yes yes

L24 yes yes L70 yes yes

L25 yes yes L71 yes yes

L26 yes yes L72 yes yes

L27 yes yes L73 T1 was not analysed yes

L29 yes yes L74 yes yes

L30 yes yes L75 yes yes

L31 yes yes L76 yes yes

L32 yes yes L77 yes yes

L33 yes yes L78 yes yes

L34 yes yes L79 yes yes

L35 yes yes L80 yes yes

L36 yes yes L81 yes yes

L37 yes yes L82 yes yes

L38 yes yes L83 yes yes

L39 yes yes L84 yes yes

L40 yes yes L85 yes yes

L41 yes yes L86 yes yes

L42 yes yes L87 yes yes

L43 T1 was not analysed yes L88 yes yes

L44 yes yes L89 yes yes

Although the questionnaire was not returned by the participant the presence of the event was inferred from the quantitative result reported

L10 did not fill in the questionnaire but reported a larger than 01 result for T1 and T2

EURL GMFF CT 0217 report

3946

Annex 4 Participants quantitative performance

The z and ζ scores of all laboratories are reported in Tables A41 and A42 and in Figures

A41 and A42 for 40-3-2 soybean in T1 and T2 respectively For consistency the

reported results are shown with two decimals the performance scores with one decimal

The z scores are displayed in green (satisfactory) orange (questionable) or red cells

(unsatisfactory)

Table A41 Quantitative results (in mm ) and performance scores of participants of CT-0217 for 40-3-2 soybean in chicken feed (T1)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 064 008 200 -019 004 003 -10 -28

L02 NRL882 057 002 200 -024 001 001 -15 -67

L03 Non-NRL 073 1520 200 -014 760 452 -04 00

L04 Non-NRL 075

-012 000 000 -03 -13

L05 NRL120 073 007 257 -014 003 002 -04 -15

L06 NRL882 075 021 200 -012 011 006 -03 -04

L07 NRL882 065 013 200 -019 007 004 -09 -19

L08 NRL120 062 023 200 -021 012 008 -11 -13

L09 NRL882 081 027 200 -009 014 007 01 01

L12 Non-NRL 100 051 200 000 026 011 10 09

L13 NRL882 094 019 173 -003 011 005 07 13

L14 Non-NRL 063

-020 000 000 -10 -50

L15 NRL882 110 011 200 004 005 002 14 46

L16 Non-NRL 117

007 000 000 17 80

L17 NRL882 072 010 200 -014 005 003 -05 -12

L18 Non-NRL 040

-040 000 000 -30 -145

L19 Non-NRL 115 063 173 006 036 014 16 11

L20 NRL120 108 032 200 003 016 006 13 19

L22 Non-NRL 051 030 223 -029 013 011 -20 -17

L23 Non-NRL 047 009 200 -033 005 004 -23 -50

L24 NRL882 072 019 200 -014 010 006 -05 -07

L25 Non-NRL 120

008 000 000 18 85

L26 NRL882 023 009 200 -064 005 008 -54 -62

L27 Non-NRL 098 015 200 -001 008 003 09 23

L29 NRL120 090 008 173 -005 005 002 05 17

L30 NRL882 045 013 200 -035 007 006 -25 -38

L31 NRL882 080 038 223 -010 017 009 00 00

L32 Non-NRL 076 039 200 -012 020 011 -02 -02

L33 NRL-120 057 014 200 -024 007 005 -15 -26

L34 NRL120 061 016 200 -021 008 006 -12 -19

L35 NRL882 097 027 200 -001 014 006 08 13

L36 Non-NRL 090 030 200 -005 015 007 05 07

L37 NRL120 065 026 200 -019 013 009 -09 -10

L38 NRL120 093 014 237 -003 006 003 07 19

L39 Non-NRL 112

005 000 000 15 71

L40 NRL882 053 008 200 -028 004 003 -18 -46

L41 NRL120 080 003 200 -010 002 001 00 00

L42 NRL882 104 031 200 002 016 006 11 17

L44 NRL120 084 007 257 -008 003 001 02 09

L45 NRL882 065 018 200 -019 009 006 -09 -14

L46 NRL882 081 018 200 -009 009 005 01 01

L47 NRL882 141 027 173 015 016 005 25 47

L48 NRL882 059 028 200 -023 014 010 -13 -13

L49 Non-NRL 089 031 200 -005 016 008 05 06

L50 NRL120 073 005 200 -014 003 001 -04 -15

L51 Non-NRL 077 049 200 -011 025 014 -02 -01

L52 Non-NRL 091 012 200 -004 006 003 06 16

L53 Non-NRL 043 081 200 -037 041 041 -27 -07

L54 NRL882 074 035 200 -013 018 010 -03 -03

L55 Non-NRL 076 019 173 -012 011 006 -02 -03

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

3246

No Answer 5 61

T1 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 71 866

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 5 61

T1 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T1

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 65 793

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 5 61

Other please specify below 1 122

No Answer 5 61

Specify the reference target(s) used (if different from above)

SOJA LEKTIN 80 bp - Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

T1 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

59 720

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

13 159

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 5 61

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T1 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 9 110

100 ng 18 220

50 ng 14 171

25 ng 6 732

15 ng 0 0

lt10 ng 0 0

No Answer 5 61

EURL GMFF CT 0217 report

3346

T1 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 3 002 4 003 1 004 6 005 7 006 2 008 4 009 7 01 37 019 1 02 1 03 1 035 1 04 1 13 1

T1 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 20 244

Determined during the in-house validation of the method 42 512

Taken from the EURL GMFF validation report 17 207

By another approach please explain below 2 244

No Answer 5 61

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

T1 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2 soybean

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 27 329

Known uncertainty of the standard method 9 110

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 6 732

No Answer 5 61

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx

Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

we used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

MON89788 soybean was detected but not quantified because results for this event were lt LOQ

We saw in the qPCR for MON89788 Ct-values at 37 - 39 but lt LOD (lt 002 )

Additional tests with one extract performed with Nucleospin Food Kit and with one extract performed with CTAB+Nucleospin Food Kit showed no significant differences in event detection and GTS 40-3-2 quantification (examined with digital PCR)

T19 not all detected GM were asked to be quantified

For T1 trace amounts of MON89788 were detected either in qualitative screening as well as in quantification The givencalculated GM (mm) of 002 is well below the known and validated LOQ of the method which is 01 We detected a low level of MON89788 but as this quantified at 0006 lower than the threshold for reporting we took this as a negative result

According course material of JRC GMO Quantification Proper calibration and Estimation of Measurement Uncertainty

MU is based on intermediate precision data

EURL GMFF CT 0217 report

3446

The sample was strongly inhibited MON89788 was detected under LOQ

In T1 the measured GM for MON89788 of 001 is below the determined LOQ (02) therefore no MU was reported

MON89788 was detected in traces at the LOD

MON89788 SOY result obtained below the LOQ but copy numbers of MON89788 was above LOD

T1 was inhibited A 10x dilution of the DNA was needed to get a reasonable result

DNA extraction of T1 has conducted using DNA extraction kits SureFood Prep Basic (S1052) and SureFood Prep Advanced (S1053) R-Biopharm AG Identification of Soya GM-lines was conducted using diagnostic kits SureFood GMO ID Roundup Ready Soya (S2030) and SureFood GMO ID RR2Y (S2034) Quantification kits were SureFood GMO Quant Roundup Ready Soya (S2014) SureFood GMO Quant 35S soya (S2028) and SureFood GMO Quant RR2Y Soya (S2029) R-Biopharm AG MON89788 soybean was detected in sample T1 but only in some cases - 2 PCR replicates were always positive out of 4 after several repetitions

Remark (precision) in relation to T18 and T19 MON89788 soybean was detected in the sample but not quantified because of lt LOQ

T2 Please select the option that applies and proceed with the questionnaire

Answers Ratio

T2 was not analysed go to Q1 0 0

T2 was analysed go to Q2 82 100

No Answer 0 0

T2 1 Why did you not analyse test item 2

Answers Ratio

a) The sample matrix is out of the scope of our laboratory 0 0

b) The methods are not validated in our laboratory 0 0

c) We could not obtain sufficient good quality DNA suitable for further analysis 0 0

d) Reference material primers probes or other reagents were not available (in time) 0 0

e) We tried but our analysis failed 0 0

f) Other practical constraints (instrument broken no personnel etc) 0 0

g) Other reason 0 0

No Answer 82 100

T2 2 Select the DNA extraction method used for T2

Answers Ratio

CTAB 40 488

NucleoSpin Food 11 134

NucleoSpin Plant 3 366

GeneSpin 4 488

Promega Wizard 3 366

DNeasy Plant 3 366

DNeasy Mericon Food 4 488

Biotecon Foodproof 5 61

SDS 4 488

Fast ID Genomic DNA 3 366

Maxwell 16 Plant DNA 0 0

Maxwell 16 Food Feed Seed 4 488

Generon Ion Force 1 122

Other 2 244

No Answer 0 0

T2 3 Select any additional DNA purification method used for T2

Answers Ratio

No additional clean-up 52 634

Additional ethanol precipitation 9 110

Eurofins DNAExtractor cleaning column 3 366

Promega Wizard DNA clean-up resin 6 732

Qiagen QIAQuick 5 61

Qiagen Genomic-Tip 20G 1 122

Other method (no need to specify) 8 976

No Answer 0 0

EURL GMFF CT 0217 report

3546

T2 4 Indicate the number of replicate DNA extractions used to obtain the results

Answers Ratio

1 0 0

2 56 683

3 9 110

4 10 122

5 2 244

6 4 488

gt6 1 122

No Answer 0 0

T2 5 Select the approach(es) used to show absence of PCR inhibition

Answers Ratio

None (no inhibition was suspected based on experience) 7 854

We run two dilutions and verify if the delta Cq or GM are as expected 32 390

We run three or four dilutions and verify if the delta Cq or GM are as expected 12 146

We perform a PCR inhibition run with a reference gene before analysis 3 or 4 dilutions linear regression extrapolation of Cq of undiluted extract compare this to the measured Cq

15 183

We add an internal positive control to the reactions and check the Cq 16 195

We verify that the amplification curves look normal 25 305

We check that the optical density ratios (OD260280 260230) are acceptable 36 439

Other 1 122

No Answer 0 0

T2 6 Select the option applicable to your analysis for 40-3-2 soybean in T2

Answers Ratio

Not tested 0 0

Detected but not quantified 2 244

Detected and quantified please fill in Q6a-6h 80 976

Found absent 0 0

No Answer 0 0

T2 7 Select the option applicable to your analysis for 68416 soybean in T2

Answers Ratio

Not tested 15 183

Detected but not quantified 1 122

Detected and quantified please fill in Q7a-7h 0 0

Found absent 66 805

No Answer 0 0

T2 8 Select the option applicable to your analysis for MON89788 soybean in T2

Answers Ratio

Not tested 5 61

Detected but not quantified 6 732

Detected and quantified please fill in Q8a-8h 3 366

Found absent 68 829

No Answer 0 0

T2 9 If applicable why did you not test or quantify all GM events in T2

Answers Ratio

a) Not applicable all GM events listed were tested and all those detected were quantified 61 744

b) The event-specific detection method is not validated in our laboratory 6 732

c) Reference material primers probes or other reagents were not available (in time) 10 122

d) The result obtained was below the LODLOQ 5 61

e) Practical constraints (instrument broken no personnel etc) 0 0

f) Other reason 2 244

No Answer 0 0

T2 6a 40-3-2 Which quantification approach was used

Answers Ratio

Standard curve method (2 calibration curves) 71 866

Delta Cq method (one calibration curve) 9 110

Digital PCR (no calibration curve) 2 244

EURL GMFF CT 0217 report

3646

No Answer 2 244

T2 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 74 902

CRM from IRMM certified in GM copy number ratio (plasmid CRM) 0 0

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 2 244

T2 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T2

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 63 768

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 6 732

Other please specify below 2 244

No Answer 5 61

Specify the reference target(s) used (if different from above)

Lectin - 74bp

lec 74bp

Lectine

SOJA LEKTIN 80 bp Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

Terry C F Harris N Event-specific detection of Roundup Ready soya using two different real time PCR detection chemistries Eur Food Res Technol (2001) 213425-431

T2 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

61 744

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

14 171

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 2 244

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T2 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 7 854

100 ng 21 256

EURL GMFF CT 0217 report

3746

50 ng 17 207

25 ng 7 854

15 ng 2 244

lt10 ng 0 0

No Answer 2 244

T2 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 5 002 6 003 1 004 7 005 7 006 2 007 1 008 2 009 8 01 39 026 1 028 1

T2 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 21 256

Determined during the in-house validation of the method 43 524

Taken from the EURL GMFF validation report 19 232

By another approach please explain below 3 366

No Answer 2 244

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample

Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit

T2 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 30 366

Known uncertainty of the standard method 8 976

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 7 854

No Answer 2 244

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

We used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

30 of the quantification result

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

Event 68416 was not tested due to non-availability of Positive control and reference material

DNA extraction of T2 has conducted using DNA extraction kits SureFood Prep Basic etc (see T1)

The NucleoSpin Food + NucleoSpin gDNA Clean-up was also used for DNA extraction from T1 and T2 samples The results for 40-3-2 were 25 - 30 lower than CTAB results T1 050 +- 018 (ww) T2 058 +- 020 (ww)

AOCS 0906-B as calibrant for MON89788 in T2 also

EURL GMFF CT 0217 report

3846

Annex 3 Qualitative performance of the participants

Table A31 Performance of the participants for the qualitative identification of GM events in CT-0217

The correct answer (yes) is that the GM event has been detected in the test item

Lab 40-3-2 detected in T1 40-3-2 detected in T2 Lab 40-3-2 detected in T1 40-3-2 detected in T2

L01 yes yes L45 yes yes

L02 yes yes L46 yes yes

L03 yes yes L47 yes yes

L04 yes yes L48 yes yes

L05 yes yes L49 yes yes

L06 yes yes L50 yes yes

L07 yes yes L51 yes yes

L08 yes yes L52 yes yes

L09 yes yes L53 yes yes

L10 yes Yes L54 yes yes

L11 yes yes L55 yes yes

L12 yes yes L56 yes yes

L13 yes yes L57 yes yes

L14 yes yes L58 yes yes

L15 yes yes L59 yes yes

L16 yes yes L60 yes yes

L17 yes yes L61 yes yes

L18 yes yes L62 yes yes

L19 yes yes L63 yes yes

L20 yes yes L66 yes yes

L22 yes yes L68 yes yes

L23 yes yes L69 yes yes

L24 yes yes L70 yes yes

L25 yes yes L71 yes yes

L26 yes yes L72 yes yes

L27 yes yes L73 T1 was not analysed yes

L29 yes yes L74 yes yes

L30 yes yes L75 yes yes

L31 yes yes L76 yes yes

L32 yes yes L77 yes yes

L33 yes yes L78 yes yes

L34 yes yes L79 yes yes

L35 yes yes L80 yes yes

L36 yes yes L81 yes yes

L37 yes yes L82 yes yes

L38 yes yes L83 yes yes

L39 yes yes L84 yes yes

L40 yes yes L85 yes yes

L41 yes yes L86 yes yes

L42 yes yes L87 yes yes

L43 T1 was not analysed yes L88 yes yes

L44 yes yes L89 yes yes

Although the questionnaire was not returned by the participant the presence of the event was inferred from the quantitative result reported

L10 did not fill in the questionnaire but reported a larger than 01 result for T1 and T2

EURL GMFF CT 0217 report

3946

Annex 4 Participants quantitative performance

The z and ζ scores of all laboratories are reported in Tables A41 and A42 and in Figures

A41 and A42 for 40-3-2 soybean in T1 and T2 respectively For consistency the

reported results are shown with two decimals the performance scores with one decimal

The z scores are displayed in green (satisfactory) orange (questionable) or red cells

(unsatisfactory)

Table A41 Quantitative results (in mm ) and performance scores of participants of CT-0217 for 40-3-2 soybean in chicken feed (T1)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 064 008 200 -019 004 003 -10 -28

L02 NRL882 057 002 200 -024 001 001 -15 -67

L03 Non-NRL 073 1520 200 -014 760 452 -04 00

L04 Non-NRL 075

-012 000 000 -03 -13

L05 NRL120 073 007 257 -014 003 002 -04 -15

L06 NRL882 075 021 200 -012 011 006 -03 -04

L07 NRL882 065 013 200 -019 007 004 -09 -19

L08 NRL120 062 023 200 -021 012 008 -11 -13

L09 NRL882 081 027 200 -009 014 007 01 01

L12 Non-NRL 100 051 200 000 026 011 10 09

L13 NRL882 094 019 173 -003 011 005 07 13

L14 Non-NRL 063

-020 000 000 -10 -50

L15 NRL882 110 011 200 004 005 002 14 46

L16 Non-NRL 117

007 000 000 17 80

L17 NRL882 072 010 200 -014 005 003 -05 -12

L18 Non-NRL 040

-040 000 000 -30 -145

L19 Non-NRL 115 063 173 006 036 014 16 11

L20 NRL120 108 032 200 003 016 006 13 19

L22 Non-NRL 051 030 223 -029 013 011 -20 -17

L23 Non-NRL 047 009 200 -033 005 004 -23 -50

L24 NRL882 072 019 200 -014 010 006 -05 -07

L25 Non-NRL 120

008 000 000 18 85

L26 NRL882 023 009 200 -064 005 008 -54 -62

L27 Non-NRL 098 015 200 -001 008 003 09 23

L29 NRL120 090 008 173 -005 005 002 05 17

L30 NRL882 045 013 200 -035 007 006 -25 -38

L31 NRL882 080 038 223 -010 017 009 00 00

L32 Non-NRL 076 039 200 -012 020 011 -02 -02

L33 NRL-120 057 014 200 -024 007 005 -15 -26

L34 NRL120 061 016 200 -021 008 006 -12 -19

L35 NRL882 097 027 200 -001 014 006 08 13

L36 Non-NRL 090 030 200 -005 015 007 05 07

L37 NRL120 065 026 200 -019 013 009 -09 -10

L38 NRL120 093 014 237 -003 006 003 07 19

L39 Non-NRL 112

005 000 000 15 71

L40 NRL882 053 008 200 -028 004 003 -18 -46

L41 NRL120 080 003 200 -010 002 001 00 00

L42 NRL882 104 031 200 002 016 006 11 17

L44 NRL120 084 007 257 -008 003 001 02 09

L45 NRL882 065 018 200 -019 009 006 -09 -14

L46 NRL882 081 018 200 -009 009 005 01 01

L47 NRL882 141 027 173 015 016 005 25 47

L48 NRL882 059 028 200 -023 014 010 -13 -13

L49 Non-NRL 089 031 200 -005 016 008 05 06

L50 NRL120 073 005 200 -014 003 001 -04 -15

L51 Non-NRL 077 049 200 -011 025 014 -02 -01

L52 Non-NRL 091 012 200 -004 006 003 06 16

L53 Non-NRL 043 081 200 -037 041 041 -27 -07

L54 NRL882 074 035 200 -013 018 010 -03 -03

L55 Non-NRL 076 019 173 -012 011 006 -02 -03

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

3346

T1 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 3 002 4 003 1 004 6 005 7 006 2 008 4 009 7 01 37 019 1 02 1 03 1 035 1 04 1 13 1

T1 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 20 244

Determined during the in-house validation of the method 42 512

Taken from the EURL GMFF validation report 17 207

By another approach please explain below 2 244

No Answer 5 61

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

T1 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2 soybean

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 27 329

Known uncertainty of the standard method 9 110

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 6 732

No Answer 5 61

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx

Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

we used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

MON89788 soybean was detected but not quantified because results for this event were lt LOQ

We saw in the qPCR for MON89788 Ct-values at 37 - 39 but lt LOD (lt 002 )

Additional tests with one extract performed with Nucleospin Food Kit and with one extract performed with CTAB+Nucleospin Food Kit showed no significant differences in event detection and GTS 40-3-2 quantification (examined with digital PCR)

T19 not all detected GM were asked to be quantified

For T1 trace amounts of MON89788 were detected either in qualitative screening as well as in quantification The givencalculated GM (mm) of 002 is well below the known and validated LOQ of the method which is 01 We detected a low level of MON89788 but as this quantified at 0006 lower than the threshold for reporting we took this as a negative result

According course material of JRC GMO Quantification Proper calibration and Estimation of Measurement Uncertainty

MU is based on intermediate precision data

EURL GMFF CT 0217 report

3446

The sample was strongly inhibited MON89788 was detected under LOQ

In T1 the measured GM for MON89788 of 001 is below the determined LOQ (02) therefore no MU was reported

MON89788 was detected in traces at the LOD

MON89788 SOY result obtained below the LOQ but copy numbers of MON89788 was above LOD

T1 was inhibited A 10x dilution of the DNA was needed to get a reasonable result

DNA extraction of T1 has conducted using DNA extraction kits SureFood Prep Basic (S1052) and SureFood Prep Advanced (S1053) R-Biopharm AG Identification of Soya GM-lines was conducted using diagnostic kits SureFood GMO ID Roundup Ready Soya (S2030) and SureFood GMO ID RR2Y (S2034) Quantification kits were SureFood GMO Quant Roundup Ready Soya (S2014) SureFood GMO Quant 35S soya (S2028) and SureFood GMO Quant RR2Y Soya (S2029) R-Biopharm AG MON89788 soybean was detected in sample T1 but only in some cases - 2 PCR replicates were always positive out of 4 after several repetitions

Remark (precision) in relation to T18 and T19 MON89788 soybean was detected in the sample but not quantified because of lt LOQ

T2 Please select the option that applies and proceed with the questionnaire

Answers Ratio

T2 was not analysed go to Q1 0 0

T2 was analysed go to Q2 82 100

No Answer 0 0

T2 1 Why did you not analyse test item 2

Answers Ratio

a) The sample matrix is out of the scope of our laboratory 0 0

b) The methods are not validated in our laboratory 0 0

c) We could not obtain sufficient good quality DNA suitable for further analysis 0 0

d) Reference material primers probes or other reagents were not available (in time) 0 0

e) We tried but our analysis failed 0 0

f) Other practical constraints (instrument broken no personnel etc) 0 0

g) Other reason 0 0

No Answer 82 100

T2 2 Select the DNA extraction method used for T2

Answers Ratio

CTAB 40 488

NucleoSpin Food 11 134

NucleoSpin Plant 3 366

GeneSpin 4 488

Promega Wizard 3 366

DNeasy Plant 3 366

DNeasy Mericon Food 4 488

Biotecon Foodproof 5 61

SDS 4 488

Fast ID Genomic DNA 3 366

Maxwell 16 Plant DNA 0 0

Maxwell 16 Food Feed Seed 4 488

Generon Ion Force 1 122

Other 2 244

No Answer 0 0

T2 3 Select any additional DNA purification method used for T2

Answers Ratio

No additional clean-up 52 634

Additional ethanol precipitation 9 110

Eurofins DNAExtractor cleaning column 3 366

Promega Wizard DNA clean-up resin 6 732

Qiagen QIAQuick 5 61

Qiagen Genomic-Tip 20G 1 122

Other method (no need to specify) 8 976

No Answer 0 0

EURL GMFF CT 0217 report

3546

T2 4 Indicate the number of replicate DNA extractions used to obtain the results

Answers Ratio

1 0 0

2 56 683

3 9 110

4 10 122

5 2 244

6 4 488

gt6 1 122

No Answer 0 0

T2 5 Select the approach(es) used to show absence of PCR inhibition

Answers Ratio

None (no inhibition was suspected based on experience) 7 854

We run two dilutions and verify if the delta Cq or GM are as expected 32 390

We run three or four dilutions and verify if the delta Cq or GM are as expected 12 146

We perform a PCR inhibition run with a reference gene before analysis 3 or 4 dilutions linear regression extrapolation of Cq of undiluted extract compare this to the measured Cq

15 183

We add an internal positive control to the reactions and check the Cq 16 195

We verify that the amplification curves look normal 25 305

We check that the optical density ratios (OD260280 260230) are acceptable 36 439

Other 1 122

No Answer 0 0

T2 6 Select the option applicable to your analysis for 40-3-2 soybean in T2

Answers Ratio

Not tested 0 0

Detected but not quantified 2 244

Detected and quantified please fill in Q6a-6h 80 976

Found absent 0 0

No Answer 0 0

T2 7 Select the option applicable to your analysis for 68416 soybean in T2

Answers Ratio

Not tested 15 183

Detected but not quantified 1 122

Detected and quantified please fill in Q7a-7h 0 0

Found absent 66 805

No Answer 0 0

T2 8 Select the option applicable to your analysis for MON89788 soybean in T2

Answers Ratio

Not tested 5 61

Detected but not quantified 6 732

Detected and quantified please fill in Q8a-8h 3 366

Found absent 68 829

No Answer 0 0

T2 9 If applicable why did you not test or quantify all GM events in T2

Answers Ratio

a) Not applicable all GM events listed were tested and all those detected were quantified 61 744

b) The event-specific detection method is not validated in our laboratory 6 732

c) Reference material primers probes or other reagents were not available (in time) 10 122

d) The result obtained was below the LODLOQ 5 61

e) Practical constraints (instrument broken no personnel etc) 0 0

f) Other reason 2 244

No Answer 0 0

T2 6a 40-3-2 Which quantification approach was used

Answers Ratio

Standard curve method (2 calibration curves) 71 866

Delta Cq method (one calibration curve) 9 110

Digital PCR (no calibration curve) 2 244

EURL GMFF CT 0217 report

3646

No Answer 2 244

T2 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 74 902

CRM from IRMM certified in GM copy number ratio (plasmid CRM) 0 0

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 2 244

T2 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T2

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 63 768

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 6 732

Other please specify below 2 244

No Answer 5 61

Specify the reference target(s) used (if different from above)

Lectin - 74bp

lec 74bp

Lectine

SOJA LEKTIN 80 bp Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

Terry C F Harris N Event-specific detection of Roundup Ready soya using two different real time PCR detection chemistries Eur Food Res Technol (2001) 213425-431

T2 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

61 744

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

14 171

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 2 244

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T2 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 7 854

100 ng 21 256

EURL GMFF CT 0217 report

3746

50 ng 17 207

25 ng 7 854

15 ng 2 244

lt10 ng 0 0

No Answer 2 244

T2 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 5 002 6 003 1 004 7 005 7 006 2 007 1 008 2 009 8 01 39 026 1 028 1

T2 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 21 256

Determined during the in-house validation of the method 43 524

Taken from the EURL GMFF validation report 19 232

By another approach please explain below 3 366

No Answer 2 244

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample

Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit

T2 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 30 366

Known uncertainty of the standard method 8 976

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 7 854

No Answer 2 244

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

We used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

30 of the quantification result

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

Event 68416 was not tested due to non-availability of Positive control and reference material

DNA extraction of T2 has conducted using DNA extraction kits SureFood Prep Basic etc (see T1)

The NucleoSpin Food + NucleoSpin gDNA Clean-up was also used for DNA extraction from T1 and T2 samples The results for 40-3-2 were 25 - 30 lower than CTAB results T1 050 +- 018 (ww) T2 058 +- 020 (ww)

AOCS 0906-B as calibrant for MON89788 in T2 also

EURL GMFF CT 0217 report

3846

Annex 3 Qualitative performance of the participants

Table A31 Performance of the participants for the qualitative identification of GM events in CT-0217

The correct answer (yes) is that the GM event has been detected in the test item

Lab 40-3-2 detected in T1 40-3-2 detected in T2 Lab 40-3-2 detected in T1 40-3-2 detected in T2

L01 yes yes L45 yes yes

L02 yes yes L46 yes yes

L03 yes yes L47 yes yes

L04 yes yes L48 yes yes

L05 yes yes L49 yes yes

L06 yes yes L50 yes yes

L07 yes yes L51 yes yes

L08 yes yes L52 yes yes

L09 yes yes L53 yes yes

L10 yes Yes L54 yes yes

L11 yes yes L55 yes yes

L12 yes yes L56 yes yes

L13 yes yes L57 yes yes

L14 yes yes L58 yes yes

L15 yes yes L59 yes yes

L16 yes yes L60 yes yes

L17 yes yes L61 yes yes

L18 yes yes L62 yes yes

L19 yes yes L63 yes yes

L20 yes yes L66 yes yes

L22 yes yes L68 yes yes

L23 yes yes L69 yes yes

L24 yes yes L70 yes yes

L25 yes yes L71 yes yes

L26 yes yes L72 yes yes

L27 yes yes L73 T1 was not analysed yes

L29 yes yes L74 yes yes

L30 yes yes L75 yes yes

L31 yes yes L76 yes yes

L32 yes yes L77 yes yes

L33 yes yes L78 yes yes

L34 yes yes L79 yes yes

L35 yes yes L80 yes yes

L36 yes yes L81 yes yes

L37 yes yes L82 yes yes

L38 yes yes L83 yes yes

L39 yes yes L84 yes yes

L40 yes yes L85 yes yes

L41 yes yes L86 yes yes

L42 yes yes L87 yes yes

L43 T1 was not analysed yes L88 yes yes

L44 yes yes L89 yes yes

Although the questionnaire was not returned by the participant the presence of the event was inferred from the quantitative result reported

L10 did not fill in the questionnaire but reported a larger than 01 result for T1 and T2

EURL GMFF CT 0217 report

3946

Annex 4 Participants quantitative performance

The z and ζ scores of all laboratories are reported in Tables A41 and A42 and in Figures

A41 and A42 for 40-3-2 soybean in T1 and T2 respectively For consistency the

reported results are shown with two decimals the performance scores with one decimal

The z scores are displayed in green (satisfactory) orange (questionable) or red cells

(unsatisfactory)

Table A41 Quantitative results (in mm ) and performance scores of participants of CT-0217 for 40-3-2 soybean in chicken feed (T1)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 064 008 200 -019 004 003 -10 -28

L02 NRL882 057 002 200 -024 001 001 -15 -67

L03 Non-NRL 073 1520 200 -014 760 452 -04 00

L04 Non-NRL 075

-012 000 000 -03 -13

L05 NRL120 073 007 257 -014 003 002 -04 -15

L06 NRL882 075 021 200 -012 011 006 -03 -04

L07 NRL882 065 013 200 -019 007 004 -09 -19

L08 NRL120 062 023 200 -021 012 008 -11 -13

L09 NRL882 081 027 200 -009 014 007 01 01

L12 Non-NRL 100 051 200 000 026 011 10 09

L13 NRL882 094 019 173 -003 011 005 07 13

L14 Non-NRL 063

-020 000 000 -10 -50

L15 NRL882 110 011 200 004 005 002 14 46

L16 Non-NRL 117

007 000 000 17 80

L17 NRL882 072 010 200 -014 005 003 -05 -12

L18 Non-NRL 040

-040 000 000 -30 -145

L19 Non-NRL 115 063 173 006 036 014 16 11

L20 NRL120 108 032 200 003 016 006 13 19

L22 Non-NRL 051 030 223 -029 013 011 -20 -17

L23 Non-NRL 047 009 200 -033 005 004 -23 -50

L24 NRL882 072 019 200 -014 010 006 -05 -07

L25 Non-NRL 120

008 000 000 18 85

L26 NRL882 023 009 200 -064 005 008 -54 -62

L27 Non-NRL 098 015 200 -001 008 003 09 23

L29 NRL120 090 008 173 -005 005 002 05 17

L30 NRL882 045 013 200 -035 007 006 -25 -38

L31 NRL882 080 038 223 -010 017 009 00 00

L32 Non-NRL 076 039 200 -012 020 011 -02 -02

L33 NRL-120 057 014 200 -024 007 005 -15 -26

L34 NRL120 061 016 200 -021 008 006 -12 -19

L35 NRL882 097 027 200 -001 014 006 08 13

L36 Non-NRL 090 030 200 -005 015 007 05 07

L37 NRL120 065 026 200 -019 013 009 -09 -10

L38 NRL120 093 014 237 -003 006 003 07 19

L39 Non-NRL 112

005 000 000 15 71

L40 NRL882 053 008 200 -028 004 003 -18 -46

L41 NRL120 080 003 200 -010 002 001 00 00

L42 NRL882 104 031 200 002 016 006 11 17

L44 NRL120 084 007 257 -008 003 001 02 09

L45 NRL882 065 018 200 -019 009 006 -09 -14

L46 NRL882 081 018 200 -009 009 005 01 01

L47 NRL882 141 027 173 015 016 005 25 47

L48 NRL882 059 028 200 -023 014 010 -13 -13

L49 Non-NRL 089 031 200 -005 016 008 05 06

L50 NRL120 073 005 200 -014 003 001 -04 -15

L51 Non-NRL 077 049 200 -011 025 014 -02 -01

L52 Non-NRL 091 012 200 -004 006 003 06 16

L53 Non-NRL 043 081 200 -037 041 041 -27 -07

L54 NRL882 074 035 200 -013 018 010 -03 -03

L55 Non-NRL 076 019 173 -012 011 006 -02 -03

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

3446

The sample was strongly inhibited MON89788 was detected under LOQ

In T1 the measured GM for MON89788 of 001 is below the determined LOQ (02) therefore no MU was reported

MON89788 was detected in traces at the LOD

MON89788 SOY result obtained below the LOQ but copy numbers of MON89788 was above LOD

T1 was inhibited A 10x dilution of the DNA was needed to get a reasonable result

DNA extraction of T1 has conducted using DNA extraction kits SureFood Prep Basic (S1052) and SureFood Prep Advanced (S1053) R-Biopharm AG Identification of Soya GM-lines was conducted using diagnostic kits SureFood GMO ID Roundup Ready Soya (S2030) and SureFood GMO ID RR2Y (S2034) Quantification kits were SureFood GMO Quant Roundup Ready Soya (S2014) SureFood GMO Quant 35S soya (S2028) and SureFood GMO Quant RR2Y Soya (S2029) R-Biopharm AG MON89788 soybean was detected in sample T1 but only in some cases - 2 PCR replicates were always positive out of 4 after several repetitions

Remark (precision) in relation to T18 and T19 MON89788 soybean was detected in the sample but not quantified because of lt LOQ

T2 Please select the option that applies and proceed with the questionnaire

Answers Ratio

T2 was not analysed go to Q1 0 0

T2 was analysed go to Q2 82 100

No Answer 0 0

T2 1 Why did you not analyse test item 2

Answers Ratio

a) The sample matrix is out of the scope of our laboratory 0 0

b) The methods are not validated in our laboratory 0 0

c) We could not obtain sufficient good quality DNA suitable for further analysis 0 0

d) Reference material primers probes or other reagents were not available (in time) 0 0

e) We tried but our analysis failed 0 0

f) Other practical constraints (instrument broken no personnel etc) 0 0

g) Other reason 0 0

No Answer 82 100

T2 2 Select the DNA extraction method used for T2

Answers Ratio

CTAB 40 488

NucleoSpin Food 11 134

NucleoSpin Plant 3 366

GeneSpin 4 488

Promega Wizard 3 366

DNeasy Plant 3 366

DNeasy Mericon Food 4 488

Biotecon Foodproof 5 61

SDS 4 488

Fast ID Genomic DNA 3 366

Maxwell 16 Plant DNA 0 0

Maxwell 16 Food Feed Seed 4 488

Generon Ion Force 1 122

Other 2 244

No Answer 0 0

T2 3 Select any additional DNA purification method used for T2

Answers Ratio

No additional clean-up 52 634

Additional ethanol precipitation 9 110

Eurofins DNAExtractor cleaning column 3 366

Promega Wizard DNA clean-up resin 6 732

Qiagen QIAQuick 5 61

Qiagen Genomic-Tip 20G 1 122

Other method (no need to specify) 8 976

No Answer 0 0

EURL GMFF CT 0217 report

3546

T2 4 Indicate the number of replicate DNA extractions used to obtain the results

Answers Ratio

1 0 0

2 56 683

3 9 110

4 10 122

5 2 244

6 4 488

gt6 1 122

No Answer 0 0

T2 5 Select the approach(es) used to show absence of PCR inhibition

Answers Ratio

None (no inhibition was suspected based on experience) 7 854

We run two dilutions and verify if the delta Cq or GM are as expected 32 390

We run three or four dilutions and verify if the delta Cq or GM are as expected 12 146

We perform a PCR inhibition run with a reference gene before analysis 3 or 4 dilutions linear regression extrapolation of Cq of undiluted extract compare this to the measured Cq

15 183

We add an internal positive control to the reactions and check the Cq 16 195

We verify that the amplification curves look normal 25 305

We check that the optical density ratios (OD260280 260230) are acceptable 36 439

Other 1 122

No Answer 0 0

T2 6 Select the option applicable to your analysis for 40-3-2 soybean in T2

Answers Ratio

Not tested 0 0

Detected but not quantified 2 244

Detected and quantified please fill in Q6a-6h 80 976

Found absent 0 0

No Answer 0 0

T2 7 Select the option applicable to your analysis for 68416 soybean in T2

Answers Ratio

Not tested 15 183

Detected but not quantified 1 122

Detected and quantified please fill in Q7a-7h 0 0

Found absent 66 805

No Answer 0 0

T2 8 Select the option applicable to your analysis for MON89788 soybean in T2

Answers Ratio

Not tested 5 61

Detected but not quantified 6 732

Detected and quantified please fill in Q8a-8h 3 366

Found absent 68 829

No Answer 0 0

T2 9 If applicable why did you not test or quantify all GM events in T2

Answers Ratio

a) Not applicable all GM events listed were tested and all those detected were quantified 61 744

b) The event-specific detection method is not validated in our laboratory 6 732

c) Reference material primers probes or other reagents were not available (in time) 10 122

d) The result obtained was below the LODLOQ 5 61

e) Practical constraints (instrument broken no personnel etc) 0 0

f) Other reason 2 244

No Answer 0 0

T2 6a 40-3-2 Which quantification approach was used

Answers Ratio

Standard curve method (2 calibration curves) 71 866

Delta Cq method (one calibration curve) 9 110

Digital PCR (no calibration curve) 2 244

EURL GMFF CT 0217 report

3646

No Answer 2 244

T2 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 74 902

CRM from IRMM certified in GM copy number ratio (plasmid CRM) 0 0

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 2 244

T2 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T2

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 63 768

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 6 732

Other please specify below 2 244

No Answer 5 61

Specify the reference target(s) used (if different from above)

Lectin - 74bp

lec 74bp

Lectine

SOJA LEKTIN 80 bp Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

Terry C F Harris N Event-specific detection of Roundup Ready soya using two different real time PCR detection chemistries Eur Food Res Technol (2001) 213425-431

T2 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

61 744

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

14 171

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 2 244

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T2 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 7 854

100 ng 21 256

EURL GMFF CT 0217 report

3746

50 ng 17 207

25 ng 7 854

15 ng 2 244

lt10 ng 0 0

No Answer 2 244

T2 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 5 002 6 003 1 004 7 005 7 006 2 007 1 008 2 009 8 01 39 026 1 028 1

T2 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 21 256

Determined during the in-house validation of the method 43 524

Taken from the EURL GMFF validation report 19 232

By another approach please explain below 3 366

No Answer 2 244

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample

Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit

T2 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 30 366

Known uncertainty of the standard method 8 976

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 7 854

No Answer 2 244

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

We used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

30 of the quantification result

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

Event 68416 was not tested due to non-availability of Positive control and reference material

DNA extraction of T2 has conducted using DNA extraction kits SureFood Prep Basic etc (see T1)

The NucleoSpin Food + NucleoSpin gDNA Clean-up was also used for DNA extraction from T1 and T2 samples The results for 40-3-2 were 25 - 30 lower than CTAB results T1 050 +- 018 (ww) T2 058 +- 020 (ww)

AOCS 0906-B as calibrant for MON89788 in T2 also

EURL GMFF CT 0217 report

3846

Annex 3 Qualitative performance of the participants

Table A31 Performance of the participants for the qualitative identification of GM events in CT-0217

The correct answer (yes) is that the GM event has been detected in the test item

Lab 40-3-2 detected in T1 40-3-2 detected in T2 Lab 40-3-2 detected in T1 40-3-2 detected in T2

L01 yes yes L45 yes yes

L02 yes yes L46 yes yes

L03 yes yes L47 yes yes

L04 yes yes L48 yes yes

L05 yes yes L49 yes yes

L06 yes yes L50 yes yes

L07 yes yes L51 yes yes

L08 yes yes L52 yes yes

L09 yes yes L53 yes yes

L10 yes Yes L54 yes yes

L11 yes yes L55 yes yes

L12 yes yes L56 yes yes

L13 yes yes L57 yes yes

L14 yes yes L58 yes yes

L15 yes yes L59 yes yes

L16 yes yes L60 yes yes

L17 yes yes L61 yes yes

L18 yes yes L62 yes yes

L19 yes yes L63 yes yes

L20 yes yes L66 yes yes

L22 yes yes L68 yes yes

L23 yes yes L69 yes yes

L24 yes yes L70 yes yes

L25 yes yes L71 yes yes

L26 yes yes L72 yes yes

L27 yes yes L73 T1 was not analysed yes

L29 yes yes L74 yes yes

L30 yes yes L75 yes yes

L31 yes yes L76 yes yes

L32 yes yes L77 yes yes

L33 yes yes L78 yes yes

L34 yes yes L79 yes yes

L35 yes yes L80 yes yes

L36 yes yes L81 yes yes

L37 yes yes L82 yes yes

L38 yes yes L83 yes yes

L39 yes yes L84 yes yes

L40 yes yes L85 yes yes

L41 yes yes L86 yes yes

L42 yes yes L87 yes yes

L43 T1 was not analysed yes L88 yes yes

L44 yes yes L89 yes yes

Although the questionnaire was not returned by the participant the presence of the event was inferred from the quantitative result reported

L10 did not fill in the questionnaire but reported a larger than 01 result for T1 and T2

EURL GMFF CT 0217 report

3946

Annex 4 Participants quantitative performance

The z and ζ scores of all laboratories are reported in Tables A41 and A42 and in Figures

A41 and A42 for 40-3-2 soybean in T1 and T2 respectively For consistency the

reported results are shown with two decimals the performance scores with one decimal

The z scores are displayed in green (satisfactory) orange (questionable) or red cells

(unsatisfactory)

Table A41 Quantitative results (in mm ) and performance scores of participants of CT-0217 for 40-3-2 soybean in chicken feed (T1)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 064 008 200 -019 004 003 -10 -28

L02 NRL882 057 002 200 -024 001 001 -15 -67

L03 Non-NRL 073 1520 200 -014 760 452 -04 00

L04 Non-NRL 075

-012 000 000 -03 -13

L05 NRL120 073 007 257 -014 003 002 -04 -15

L06 NRL882 075 021 200 -012 011 006 -03 -04

L07 NRL882 065 013 200 -019 007 004 -09 -19

L08 NRL120 062 023 200 -021 012 008 -11 -13

L09 NRL882 081 027 200 -009 014 007 01 01

L12 Non-NRL 100 051 200 000 026 011 10 09

L13 NRL882 094 019 173 -003 011 005 07 13

L14 Non-NRL 063

-020 000 000 -10 -50

L15 NRL882 110 011 200 004 005 002 14 46

L16 Non-NRL 117

007 000 000 17 80

L17 NRL882 072 010 200 -014 005 003 -05 -12

L18 Non-NRL 040

-040 000 000 -30 -145

L19 Non-NRL 115 063 173 006 036 014 16 11

L20 NRL120 108 032 200 003 016 006 13 19

L22 Non-NRL 051 030 223 -029 013 011 -20 -17

L23 Non-NRL 047 009 200 -033 005 004 -23 -50

L24 NRL882 072 019 200 -014 010 006 -05 -07

L25 Non-NRL 120

008 000 000 18 85

L26 NRL882 023 009 200 -064 005 008 -54 -62

L27 Non-NRL 098 015 200 -001 008 003 09 23

L29 NRL120 090 008 173 -005 005 002 05 17

L30 NRL882 045 013 200 -035 007 006 -25 -38

L31 NRL882 080 038 223 -010 017 009 00 00

L32 Non-NRL 076 039 200 -012 020 011 -02 -02

L33 NRL-120 057 014 200 -024 007 005 -15 -26

L34 NRL120 061 016 200 -021 008 006 -12 -19

L35 NRL882 097 027 200 -001 014 006 08 13

L36 Non-NRL 090 030 200 -005 015 007 05 07

L37 NRL120 065 026 200 -019 013 009 -09 -10

L38 NRL120 093 014 237 -003 006 003 07 19

L39 Non-NRL 112

005 000 000 15 71

L40 NRL882 053 008 200 -028 004 003 -18 -46

L41 NRL120 080 003 200 -010 002 001 00 00

L42 NRL882 104 031 200 002 016 006 11 17

L44 NRL120 084 007 257 -008 003 001 02 09

L45 NRL882 065 018 200 -019 009 006 -09 -14

L46 NRL882 081 018 200 -009 009 005 01 01

L47 NRL882 141 027 173 015 016 005 25 47

L48 NRL882 059 028 200 -023 014 010 -13 -13

L49 Non-NRL 089 031 200 -005 016 008 05 06

L50 NRL120 073 005 200 -014 003 001 -04 -15

L51 Non-NRL 077 049 200 -011 025 014 -02 -01

L52 Non-NRL 091 012 200 -004 006 003 06 16

L53 Non-NRL 043 081 200 -037 041 041 -27 -07

L54 NRL882 074 035 200 -013 018 010 -03 -03

L55 Non-NRL 076 019 173 -012 011 006 -02 -03

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

3546

T2 4 Indicate the number of replicate DNA extractions used to obtain the results

Answers Ratio

1 0 0

2 56 683

3 9 110

4 10 122

5 2 244

6 4 488

gt6 1 122

No Answer 0 0

T2 5 Select the approach(es) used to show absence of PCR inhibition

Answers Ratio

None (no inhibition was suspected based on experience) 7 854

We run two dilutions and verify if the delta Cq or GM are as expected 32 390

We run three or four dilutions and verify if the delta Cq or GM are as expected 12 146

We perform a PCR inhibition run with a reference gene before analysis 3 or 4 dilutions linear regression extrapolation of Cq of undiluted extract compare this to the measured Cq

15 183

We add an internal positive control to the reactions and check the Cq 16 195

We verify that the amplification curves look normal 25 305

We check that the optical density ratios (OD260280 260230) are acceptable 36 439

Other 1 122

No Answer 0 0

T2 6 Select the option applicable to your analysis for 40-3-2 soybean in T2

Answers Ratio

Not tested 0 0

Detected but not quantified 2 244

Detected and quantified please fill in Q6a-6h 80 976

Found absent 0 0

No Answer 0 0

T2 7 Select the option applicable to your analysis for 68416 soybean in T2

Answers Ratio

Not tested 15 183

Detected but not quantified 1 122

Detected and quantified please fill in Q7a-7h 0 0

Found absent 66 805

No Answer 0 0

T2 8 Select the option applicable to your analysis for MON89788 soybean in T2

Answers Ratio

Not tested 5 61

Detected but not quantified 6 732

Detected and quantified please fill in Q8a-8h 3 366

Found absent 68 829

No Answer 0 0

T2 9 If applicable why did you not test or quantify all GM events in T2

Answers Ratio

a) Not applicable all GM events listed were tested and all those detected were quantified 61 744

b) The event-specific detection method is not validated in our laboratory 6 732

c) Reference material primers probes or other reagents were not available (in time) 10 122

d) The result obtained was below the LODLOQ 5 61

e) Practical constraints (instrument broken no personnel etc) 0 0

f) Other reason 2 244

No Answer 0 0

T2 6a 40-3-2 Which quantification approach was used

Answers Ratio

Standard curve method (2 calibration curves) 71 866

Delta Cq method (one calibration curve) 9 110

Digital PCR (no calibration curve) 2 244

EURL GMFF CT 0217 report

3646

No Answer 2 244

T2 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 74 902

CRM from IRMM certified in GM copy number ratio (plasmid CRM) 0 0

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 2 244

T2 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T2

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 63 768

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 6 732

Other please specify below 2 244

No Answer 5 61

Specify the reference target(s) used (if different from above)

Lectin - 74bp

lec 74bp

Lectine

SOJA LEKTIN 80 bp Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

Terry C F Harris N Event-specific detection of Roundup Ready soya using two different real time PCR detection chemistries Eur Food Res Technol (2001) 213425-431

T2 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

61 744

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

14 171

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 2 244

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T2 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 7 854

100 ng 21 256

EURL GMFF CT 0217 report

3746

50 ng 17 207

25 ng 7 854

15 ng 2 244

lt10 ng 0 0

No Answer 2 244

T2 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 5 002 6 003 1 004 7 005 7 006 2 007 1 008 2 009 8 01 39 026 1 028 1

T2 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 21 256

Determined during the in-house validation of the method 43 524

Taken from the EURL GMFF validation report 19 232

By another approach please explain below 3 366

No Answer 2 244

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample

Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit

T2 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 30 366

Known uncertainty of the standard method 8 976

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 7 854

No Answer 2 244

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

We used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

30 of the quantification result

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

Event 68416 was not tested due to non-availability of Positive control and reference material

DNA extraction of T2 has conducted using DNA extraction kits SureFood Prep Basic etc (see T1)

The NucleoSpin Food + NucleoSpin gDNA Clean-up was also used for DNA extraction from T1 and T2 samples The results for 40-3-2 were 25 - 30 lower than CTAB results T1 050 +- 018 (ww) T2 058 +- 020 (ww)

AOCS 0906-B as calibrant for MON89788 in T2 also

EURL GMFF CT 0217 report

3846

Annex 3 Qualitative performance of the participants

Table A31 Performance of the participants for the qualitative identification of GM events in CT-0217

The correct answer (yes) is that the GM event has been detected in the test item

Lab 40-3-2 detected in T1 40-3-2 detected in T2 Lab 40-3-2 detected in T1 40-3-2 detected in T2

L01 yes yes L45 yes yes

L02 yes yes L46 yes yes

L03 yes yes L47 yes yes

L04 yes yes L48 yes yes

L05 yes yes L49 yes yes

L06 yes yes L50 yes yes

L07 yes yes L51 yes yes

L08 yes yes L52 yes yes

L09 yes yes L53 yes yes

L10 yes Yes L54 yes yes

L11 yes yes L55 yes yes

L12 yes yes L56 yes yes

L13 yes yes L57 yes yes

L14 yes yes L58 yes yes

L15 yes yes L59 yes yes

L16 yes yes L60 yes yes

L17 yes yes L61 yes yes

L18 yes yes L62 yes yes

L19 yes yes L63 yes yes

L20 yes yes L66 yes yes

L22 yes yes L68 yes yes

L23 yes yes L69 yes yes

L24 yes yes L70 yes yes

L25 yes yes L71 yes yes

L26 yes yes L72 yes yes

L27 yes yes L73 T1 was not analysed yes

L29 yes yes L74 yes yes

L30 yes yes L75 yes yes

L31 yes yes L76 yes yes

L32 yes yes L77 yes yes

L33 yes yes L78 yes yes

L34 yes yes L79 yes yes

L35 yes yes L80 yes yes

L36 yes yes L81 yes yes

L37 yes yes L82 yes yes

L38 yes yes L83 yes yes

L39 yes yes L84 yes yes

L40 yes yes L85 yes yes

L41 yes yes L86 yes yes

L42 yes yes L87 yes yes

L43 T1 was not analysed yes L88 yes yes

L44 yes yes L89 yes yes

Although the questionnaire was not returned by the participant the presence of the event was inferred from the quantitative result reported

L10 did not fill in the questionnaire but reported a larger than 01 result for T1 and T2

EURL GMFF CT 0217 report

3946

Annex 4 Participants quantitative performance

The z and ζ scores of all laboratories are reported in Tables A41 and A42 and in Figures

A41 and A42 for 40-3-2 soybean in T1 and T2 respectively For consistency the

reported results are shown with two decimals the performance scores with one decimal

The z scores are displayed in green (satisfactory) orange (questionable) or red cells

(unsatisfactory)

Table A41 Quantitative results (in mm ) and performance scores of participants of CT-0217 for 40-3-2 soybean in chicken feed (T1)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 064 008 200 -019 004 003 -10 -28

L02 NRL882 057 002 200 -024 001 001 -15 -67

L03 Non-NRL 073 1520 200 -014 760 452 -04 00

L04 Non-NRL 075

-012 000 000 -03 -13

L05 NRL120 073 007 257 -014 003 002 -04 -15

L06 NRL882 075 021 200 -012 011 006 -03 -04

L07 NRL882 065 013 200 -019 007 004 -09 -19

L08 NRL120 062 023 200 -021 012 008 -11 -13

L09 NRL882 081 027 200 -009 014 007 01 01

L12 Non-NRL 100 051 200 000 026 011 10 09

L13 NRL882 094 019 173 -003 011 005 07 13

L14 Non-NRL 063

-020 000 000 -10 -50

L15 NRL882 110 011 200 004 005 002 14 46

L16 Non-NRL 117

007 000 000 17 80

L17 NRL882 072 010 200 -014 005 003 -05 -12

L18 Non-NRL 040

-040 000 000 -30 -145

L19 Non-NRL 115 063 173 006 036 014 16 11

L20 NRL120 108 032 200 003 016 006 13 19

L22 Non-NRL 051 030 223 -029 013 011 -20 -17

L23 Non-NRL 047 009 200 -033 005 004 -23 -50

L24 NRL882 072 019 200 -014 010 006 -05 -07

L25 Non-NRL 120

008 000 000 18 85

L26 NRL882 023 009 200 -064 005 008 -54 -62

L27 Non-NRL 098 015 200 -001 008 003 09 23

L29 NRL120 090 008 173 -005 005 002 05 17

L30 NRL882 045 013 200 -035 007 006 -25 -38

L31 NRL882 080 038 223 -010 017 009 00 00

L32 Non-NRL 076 039 200 -012 020 011 -02 -02

L33 NRL-120 057 014 200 -024 007 005 -15 -26

L34 NRL120 061 016 200 -021 008 006 -12 -19

L35 NRL882 097 027 200 -001 014 006 08 13

L36 Non-NRL 090 030 200 -005 015 007 05 07

L37 NRL120 065 026 200 -019 013 009 -09 -10

L38 NRL120 093 014 237 -003 006 003 07 19

L39 Non-NRL 112

005 000 000 15 71

L40 NRL882 053 008 200 -028 004 003 -18 -46

L41 NRL120 080 003 200 -010 002 001 00 00

L42 NRL882 104 031 200 002 016 006 11 17

L44 NRL120 084 007 257 -008 003 001 02 09

L45 NRL882 065 018 200 -019 009 006 -09 -14

L46 NRL882 081 018 200 -009 009 005 01 01

L47 NRL882 141 027 173 015 016 005 25 47

L48 NRL882 059 028 200 -023 014 010 -13 -13

L49 Non-NRL 089 031 200 -005 016 008 05 06

L50 NRL120 073 005 200 -014 003 001 -04 -15

L51 Non-NRL 077 049 200 -011 025 014 -02 -01

L52 Non-NRL 091 012 200 -004 006 003 06 16

L53 Non-NRL 043 081 200 -037 041 041 -27 -07

L54 NRL882 074 035 200 -013 018 010 -03 -03

L55 Non-NRL 076 019 173 -012 011 006 -02 -03

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

3646

No Answer 2 244

T2 6b Select the calibrant used for the 40-3-2 standard curve

Answers Ratio

CRM from JRC (ex-IRMM) certified in GM mass fraction (gkg) 74 902

CRM from IRMM certified in GM copy number ratio (plasmid CRM) 0 0

Non-certified RM (eg lab QC material) expressed in GM mass fraction (gkg or mm ) 1 122

Non-certified RM (eg lab QC material) expressed in GM DNA copy number ratio (eg determined by digital PCR)

4 488

No calibrant used digital PCR done 1 122

No Answer 2 244

T2 6c Select the endogenous target(s) used for relative quantification of 40-3-2 soybean in T2

Answers Ratio

Soybean lec 74 bp (40-3-2 MON89788 MON87701 44406 356043 305423 etc) 63 768

Soybean lec 81 bp (Pauli et al 2001) 5 61

Soybean lec 102 bp (A5547 FG72) 1 122

Soybean lec 105 bp (A2704) 0 0

Soybean lec 118 bp (Shindo et al 2002) 6 732

Other please specify below 2 244

No Answer 5 61

Specify the reference target(s) used (if different from above)

Lectin - 74bp

lec 74bp

Lectine

SOJA LEKTIN 80 bp Va M Pijnenburrg H Gendre F Brignon P (1999) J Agric Food Chem 475261-5266

Terry C F Harris N Event-specific detection of Roundup Ready soya using two different real time PCR detection chemistries Eur Food Res Technol (2001) 213425-431

T2 6d Clarify the unit of measurement used and any conversion between units if applicable Carefully read the choices

below and select the one used in the measurements that resulted in a final result in GM mm for 40-3-2 If unclear or a

different approach was used please clarify this in the free text box below

Answers Ratio

The RM and the calibration standards were expressed in mass (or mass ) no conversion factor was applied

61 744

The calibration standards were expressed in DNA copies calculated from the RM in gkg but a conversion factor of 1 was applied (eg 10 mm GM = 10 cpcp GM corresponding to a 10x dilution of a 100 RM)

14 171

The calibration standards were expressed in DNA copies calculated from the RM in gkg and a conversion factor gt1 was applied to take account of the zygosity and target gene copies (double conversion applied) a conversion factor (eg 2) was used to convert from mass to copies (eg 20 mm GM = 10 cpcp GM corresponding to a 5x dilution of a 100 RM) the final result was again converted to mm by using the same conversion factor (eg x 2) Please specify this factor below

0 0

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) A conversion factor was applied onto the final GM please specify this factor below

2 244

The measurements were done in DNA copies (as the RM used was expressed in this unit or digital PCR was used) No conversion factor was applied onto the final GM

3 366

No Answer 2 244

Conversion factor used to turn results into mm if applicable andor clarification on preparation of standards

1

Conversion factor of 1

T2 6e What was the amount of sample DNA (ng) used per PCR for 40-3-2 Choose the concentration that is closest to what

you used If applicable select multiple concentrations (eg if several dilutions were tested) but only those of which the

result was used to determine the reported GM

Answers Ratio

DNA concentration not determined 12 146

250 ng 5 61

200 ng 30 366

150 ng 7 854

100 ng 21 256

EURL GMFF CT 0217 report

3746

50 ng 17 207

25 ng 7 854

15 ng 2 244

lt10 ng 0 0

No Answer 2 244

T2 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 5 002 6 003 1 004 7 005 7 006 2 007 1 008 2 009 8 01 39 026 1 028 1

T2 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 21 256

Determined during the in-house validation of the method 43 524

Taken from the EURL GMFF validation report 19 232

By another approach please explain below 3 366

No Answer 2 244

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample

Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit

T2 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 30 366

Known uncertainty of the standard method 8 976

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 7 854

No Answer 2 244

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

We used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

30 of the quantification result

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

Event 68416 was not tested due to non-availability of Positive control and reference material

DNA extraction of T2 has conducted using DNA extraction kits SureFood Prep Basic etc (see T1)

The NucleoSpin Food + NucleoSpin gDNA Clean-up was also used for DNA extraction from T1 and T2 samples The results for 40-3-2 were 25 - 30 lower than CTAB results T1 050 +- 018 (ww) T2 058 +- 020 (ww)

AOCS 0906-B as calibrant for MON89788 in T2 also

EURL GMFF CT 0217 report

3846

Annex 3 Qualitative performance of the participants

Table A31 Performance of the participants for the qualitative identification of GM events in CT-0217

The correct answer (yes) is that the GM event has been detected in the test item

Lab 40-3-2 detected in T1 40-3-2 detected in T2 Lab 40-3-2 detected in T1 40-3-2 detected in T2

L01 yes yes L45 yes yes

L02 yes yes L46 yes yes

L03 yes yes L47 yes yes

L04 yes yes L48 yes yes

L05 yes yes L49 yes yes

L06 yes yes L50 yes yes

L07 yes yes L51 yes yes

L08 yes yes L52 yes yes

L09 yes yes L53 yes yes

L10 yes Yes L54 yes yes

L11 yes yes L55 yes yes

L12 yes yes L56 yes yes

L13 yes yes L57 yes yes

L14 yes yes L58 yes yes

L15 yes yes L59 yes yes

L16 yes yes L60 yes yes

L17 yes yes L61 yes yes

L18 yes yes L62 yes yes

L19 yes yes L63 yes yes

L20 yes yes L66 yes yes

L22 yes yes L68 yes yes

L23 yes yes L69 yes yes

L24 yes yes L70 yes yes

L25 yes yes L71 yes yes

L26 yes yes L72 yes yes

L27 yes yes L73 T1 was not analysed yes

L29 yes yes L74 yes yes

L30 yes yes L75 yes yes

L31 yes yes L76 yes yes

L32 yes yes L77 yes yes

L33 yes yes L78 yes yes

L34 yes yes L79 yes yes

L35 yes yes L80 yes yes

L36 yes yes L81 yes yes

L37 yes yes L82 yes yes

L38 yes yes L83 yes yes

L39 yes yes L84 yes yes

L40 yes yes L85 yes yes

L41 yes yes L86 yes yes

L42 yes yes L87 yes yes

L43 T1 was not analysed yes L88 yes yes

L44 yes yes L89 yes yes

Although the questionnaire was not returned by the participant the presence of the event was inferred from the quantitative result reported

L10 did not fill in the questionnaire but reported a larger than 01 result for T1 and T2

EURL GMFF CT 0217 report

3946

Annex 4 Participants quantitative performance

The z and ζ scores of all laboratories are reported in Tables A41 and A42 and in Figures

A41 and A42 for 40-3-2 soybean in T1 and T2 respectively For consistency the

reported results are shown with two decimals the performance scores with one decimal

The z scores are displayed in green (satisfactory) orange (questionable) or red cells

(unsatisfactory)

Table A41 Quantitative results (in mm ) and performance scores of participants of CT-0217 for 40-3-2 soybean in chicken feed (T1)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 064 008 200 -019 004 003 -10 -28

L02 NRL882 057 002 200 -024 001 001 -15 -67

L03 Non-NRL 073 1520 200 -014 760 452 -04 00

L04 Non-NRL 075

-012 000 000 -03 -13

L05 NRL120 073 007 257 -014 003 002 -04 -15

L06 NRL882 075 021 200 -012 011 006 -03 -04

L07 NRL882 065 013 200 -019 007 004 -09 -19

L08 NRL120 062 023 200 -021 012 008 -11 -13

L09 NRL882 081 027 200 -009 014 007 01 01

L12 Non-NRL 100 051 200 000 026 011 10 09

L13 NRL882 094 019 173 -003 011 005 07 13

L14 Non-NRL 063

-020 000 000 -10 -50

L15 NRL882 110 011 200 004 005 002 14 46

L16 Non-NRL 117

007 000 000 17 80

L17 NRL882 072 010 200 -014 005 003 -05 -12

L18 Non-NRL 040

-040 000 000 -30 -145

L19 Non-NRL 115 063 173 006 036 014 16 11

L20 NRL120 108 032 200 003 016 006 13 19

L22 Non-NRL 051 030 223 -029 013 011 -20 -17

L23 Non-NRL 047 009 200 -033 005 004 -23 -50

L24 NRL882 072 019 200 -014 010 006 -05 -07

L25 Non-NRL 120

008 000 000 18 85

L26 NRL882 023 009 200 -064 005 008 -54 -62

L27 Non-NRL 098 015 200 -001 008 003 09 23

L29 NRL120 090 008 173 -005 005 002 05 17

L30 NRL882 045 013 200 -035 007 006 -25 -38

L31 NRL882 080 038 223 -010 017 009 00 00

L32 Non-NRL 076 039 200 -012 020 011 -02 -02

L33 NRL-120 057 014 200 -024 007 005 -15 -26

L34 NRL120 061 016 200 -021 008 006 -12 -19

L35 NRL882 097 027 200 -001 014 006 08 13

L36 Non-NRL 090 030 200 -005 015 007 05 07

L37 NRL120 065 026 200 -019 013 009 -09 -10

L38 NRL120 093 014 237 -003 006 003 07 19

L39 Non-NRL 112

005 000 000 15 71

L40 NRL882 053 008 200 -028 004 003 -18 -46

L41 NRL120 080 003 200 -010 002 001 00 00

L42 NRL882 104 031 200 002 016 006 11 17

L44 NRL120 084 007 257 -008 003 001 02 09

L45 NRL882 065 018 200 -019 009 006 -09 -14

L46 NRL882 081 018 200 -009 009 005 01 01

L47 NRL882 141 027 173 015 016 005 25 47

L48 NRL882 059 028 200 -023 014 010 -13 -13

L49 Non-NRL 089 031 200 -005 016 008 05 06

L50 NRL120 073 005 200 -014 003 001 -04 -15

L51 Non-NRL 077 049 200 -011 025 014 -02 -01

L52 Non-NRL 091 012 200 -004 006 003 06 16

L53 Non-NRL 043 081 200 -037 041 041 -27 -07

L54 NRL882 074 035 200 -013 018 010 -03 -03

L55 Non-NRL 076 019 173 -012 011 006 -02 -03

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

3746

50 ng 17 207

25 ng 7 854

15 ng 2 244

lt10 ng 0 0

No Answer 2 244

T2 6f What was the LOQ (in mm ) for the 40-3-2 quantification

Answers

001 5 002 6 003 1 004 7 005 7 006 2 007 1 008 2 009 8 01 39 026 1 028 1

T2 6g How was the LOQ for 40-3-2 determined (if applicable)

Answers Ratio

Determined from the qPCR analysis for the current sample 21 256

Determined during the in-house validation of the method 43 524

Taken from the EURL GMFF validation report 19 232

By another approach please explain below 3 366

No Answer 2 244

Explanation on alternative LOQ determination

Determined from the digital PCR analysis for the current sample

Information about LOQ introduced by the manufacturer (R-Biopharm) of diagnostic kit

T2 6h How did you estimate the measurement uncertainty on the result reported for 40-3-2

Answers Ratio

Uncertainty budget (ISO GUM) 2 244

Uncertainty of the method (in-house validation) 30 366

Known uncertainty of the standard method 8 976

Measurement of replicates (precision) 32 390

From interlaboratory comparison data 4 488

Estimation based on judgement 2 244

In another way please specify below 7 854

No Answer 2 244

Explanation on alternative determination of measurement uncertainty

U=Sa radic1p+1n+(c0-c)2Sxx Uncertainty=Coverage Factor (P=95 anf f=n-1) Standard Deviation Square-root (Number of Measurements)

MU was estimated according to the Guidance Document on MU for GMO Testing Laboratories JRC ISSN 1018-5593

We used a calculated k-factor based on the number of repeats

Estimation based on within laboratory reproducibility

95 confidence interval of the results for the current sample

30 of the quantification result

The u was obtain through the estimation of the sd taking into account the repeatability and intermediate precision associated with the test sample Coverage factor k = 2 was applied then

Combined Uncertainty (CRM + measurement) following Application note (Lingsinger 2005 JRC Geel)

Additional comments and suggestions

Event 68416 was not tested due to non-availability of Positive control and reference material

DNA extraction of T2 has conducted using DNA extraction kits SureFood Prep Basic etc (see T1)

The NucleoSpin Food + NucleoSpin gDNA Clean-up was also used for DNA extraction from T1 and T2 samples The results for 40-3-2 were 25 - 30 lower than CTAB results T1 050 +- 018 (ww) T2 058 +- 020 (ww)

AOCS 0906-B as calibrant for MON89788 in T2 also

EURL GMFF CT 0217 report

3846

Annex 3 Qualitative performance of the participants

Table A31 Performance of the participants for the qualitative identification of GM events in CT-0217

The correct answer (yes) is that the GM event has been detected in the test item

Lab 40-3-2 detected in T1 40-3-2 detected in T2 Lab 40-3-2 detected in T1 40-3-2 detected in T2

L01 yes yes L45 yes yes

L02 yes yes L46 yes yes

L03 yes yes L47 yes yes

L04 yes yes L48 yes yes

L05 yes yes L49 yes yes

L06 yes yes L50 yes yes

L07 yes yes L51 yes yes

L08 yes yes L52 yes yes

L09 yes yes L53 yes yes

L10 yes Yes L54 yes yes

L11 yes yes L55 yes yes

L12 yes yes L56 yes yes

L13 yes yes L57 yes yes

L14 yes yes L58 yes yes

L15 yes yes L59 yes yes

L16 yes yes L60 yes yes

L17 yes yes L61 yes yes

L18 yes yes L62 yes yes

L19 yes yes L63 yes yes

L20 yes yes L66 yes yes

L22 yes yes L68 yes yes

L23 yes yes L69 yes yes

L24 yes yes L70 yes yes

L25 yes yes L71 yes yes

L26 yes yes L72 yes yes

L27 yes yes L73 T1 was not analysed yes

L29 yes yes L74 yes yes

L30 yes yes L75 yes yes

L31 yes yes L76 yes yes

L32 yes yes L77 yes yes

L33 yes yes L78 yes yes

L34 yes yes L79 yes yes

L35 yes yes L80 yes yes

L36 yes yes L81 yes yes

L37 yes yes L82 yes yes

L38 yes yes L83 yes yes

L39 yes yes L84 yes yes

L40 yes yes L85 yes yes

L41 yes yes L86 yes yes

L42 yes yes L87 yes yes

L43 T1 was not analysed yes L88 yes yes

L44 yes yes L89 yes yes

Although the questionnaire was not returned by the participant the presence of the event was inferred from the quantitative result reported

L10 did not fill in the questionnaire but reported a larger than 01 result for T1 and T2

EURL GMFF CT 0217 report

3946

Annex 4 Participants quantitative performance

The z and ζ scores of all laboratories are reported in Tables A41 and A42 and in Figures

A41 and A42 for 40-3-2 soybean in T1 and T2 respectively For consistency the

reported results are shown with two decimals the performance scores with one decimal

The z scores are displayed in green (satisfactory) orange (questionable) or red cells

(unsatisfactory)

Table A41 Quantitative results (in mm ) and performance scores of participants of CT-0217 for 40-3-2 soybean in chicken feed (T1)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 064 008 200 -019 004 003 -10 -28

L02 NRL882 057 002 200 -024 001 001 -15 -67

L03 Non-NRL 073 1520 200 -014 760 452 -04 00

L04 Non-NRL 075

-012 000 000 -03 -13

L05 NRL120 073 007 257 -014 003 002 -04 -15

L06 NRL882 075 021 200 -012 011 006 -03 -04

L07 NRL882 065 013 200 -019 007 004 -09 -19

L08 NRL120 062 023 200 -021 012 008 -11 -13

L09 NRL882 081 027 200 -009 014 007 01 01

L12 Non-NRL 100 051 200 000 026 011 10 09

L13 NRL882 094 019 173 -003 011 005 07 13

L14 Non-NRL 063

-020 000 000 -10 -50

L15 NRL882 110 011 200 004 005 002 14 46

L16 Non-NRL 117

007 000 000 17 80

L17 NRL882 072 010 200 -014 005 003 -05 -12

L18 Non-NRL 040

-040 000 000 -30 -145

L19 Non-NRL 115 063 173 006 036 014 16 11

L20 NRL120 108 032 200 003 016 006 13 19

L22 Non-NRL 051 030 223 -029 013 011 -20 -17

L23 Non-NRL 047 009 200 -033 005 004 -23 -50

L24 NRL882 072 019 200 -014 010 006 -05 -07

L25 Non-NRL 120

008 000 000 18 85

L26 NRL882 023 009 200 -064 005 008 -54 -62

L27 Non-NRL 098 015 200 -001 008 003 09 23

L29 NRL120 090 008 173 -005 005 002 05 17

L30 NRL882 045 013 200 -035 007 006 -25 -38

L31 NRL882 080 038 223 -010 017 009 00 00

L32 Non-NRL 076 039 200 -012 020 011 -02 -02

L33 NRL-120 057 014 200 -024 007 005 -15 -26

L34 NRL120 061 016 200 -021 008 006 -12 -19

L35 NRL882 097 027 200 -001 014 006 08 13

L36 Non-NRL 090 030 200 -005 015 007 05 07

L37 NRL120 065 026 200 -019 013 009 -09 -10

L38 NRL120 093 014 237 -003 006 003 07 19

L39 Non-NRL 112

005 000 000 15 71

L40 NRL882 053 008 200 -028 004 003 -18 -46

L41 NRL120 080 003 200 -010 002 001 00 00

L42 NRL882 104 031 200 002 016 006 11 17

L44 NRL120 084 007 257 -008 003 001 02 09

L45 NRL882 065 018 200 -019 009 006 -09 -14

L46 NRL882 081 018 200 -009 009 005 01 01

L47 NRL882 141 027 173 015 016 005 25 47

L48 NRL882 059 028 200 -023 014 010 -13 -13

L49 Non-NRL 089 031 200 -005 016 008 05 06

L50 NRL120 073 005 200 -014 003 001 -04 -15

L51 Non-NRL 077 049 200 -011 025 014 -02 -01

L52 Non-NRL 091 012 200 -004 006 003 06 16

L53 Non-NRL 043 081 200 -037 041 041 -27 -07

L54 NRL882 074 035 200 -013 018 010 -03 -03

L55 Non-NRL 076 019 173 -012 011 006 -02 -03

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

3846

Annex 3 Qualitative performance of the participants

Table A31 Performance of the participants for the qualitative identification of GM events in CT-0217

The correct answer (yes) is that the GM event has been detected in the test item

Lab 40-3-2 detected in T1 40-3-2 detected in T2 Lab 40-3-2 detected in T1 40-3-2 detected in T2

L01 yes yes L45 yes yes

L02 yes yes L46 yes yes

L03 yes yes L47 yes yes

L04 yes yes L48 yes yes

L05 yes yes L49 yes yes

L06 yes yes L50 yes yes

L07 yes yes L51 yes yes

L08 yes yes L52 yes yes

L09 yes yes L53 yes yes

L10 yes Yes L54 yes yes

L11 yes yes L55 yes yes

L12 yes yes L56 yes yes

L13 yes yes L57 yes yes

L14 yes yes L58 yes yes

L15 yes yes L59 yes yes

L16 yes yes L60 yes yes

L17 yes yes L61 yes yes

L18 yes yes L62 yes yes

L19 yes yes L63 yes yes

L20 yes yes L66 yes yes

L22 yes yes L68 yes yes

L23 yes yes L69 yes yes

L24 yes yes L70 yes yes

L25 yes yes L71 yes yes

L26 yes yes L72 yes yes

L27 yes yes L73 T1 was not analysed yes

L29 yes yes L74 yes yes

L30 yes yes L75 yes yes

L31 yes yes L76 yes yes

L32 yes yes L77 yes yes

L33 yes yes L78 yes yes

L34 yes yes L79 yes yes

L35 yes yes L80 yes yes

L36 yes yes L81 yes yes

L37 yes yes L82 yes yes

L38 yes yes L83 yes yes

L39 yes yes L84 yes yes

L40 yes yes L85 yes yes

L41 yes yes L86 yes yes

L42 yes yes L87 yes yes

L43 T1 was not analysed yes L88 yes yes

L44 yes yes L89 yes yes

Although the questionnaire was not returned by the participant the presence of the event was inferred from the quantitative result reported

L10 did not fill in the questionnaire but reported a larger than 01 result for T1 and T2

EURL GMFF CT 0217 report

3946

Annex 4 Participants quantitative performance

The z and ζ scores of all laboratories are reported in Tables A41 and A42 and in Figures

A41 and A42 for 40-3-2 soybean in T1 and T2 respectively For consistency the

reported results are shown with two decimals the performance scores with one decimal

The z scores are displayed in green (satisfactory) orange (questionable) or red cells

(unsatisfactory)

Table A41 Quantitative results (in mm ) and performance scores of participants of CT-0217 for 40-3-2 soybean in chicken feed (T1)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 064 008 200 -019 004 003 -10 -28

L02 NRL882 057 002 200 -024 001 001 -15 -67

L03 Non-NRL 073 1520 200 -014 760 452 -04 00

L04 Non-NRL 075

-012 000 000 -03 -13

L05 NRL120 073 007 257 -014 003 002 -04 -15

L06 NRL882 075 021 200 -012 011 006 -03 -04

L07 NRL882 065 013 200 -019 007 004 -09 -19

L08 NRL120 062 023 200 -021 012 008 -11 -13

L09 NRL882 081 027 200 -009 014 007 01 01

L12 Non-NRL 100 051 200 000 026 011 10 09

L13 NRL882 094 019 173 -003 011 005 07 13

L14 Non-NRL 063

-020 000 000 -10 -50

L15 NRL882 110 011 200 004 005 002 14 46

L16 Non-NRL 117

007 000 000 17 80

L17 NRL882 072 010 200 -014 005 003 -05 -12

L18 Non-NRL 040

-040 000 000 -30 -145

L19 Non-NRL 115 063 173 006 036 014 16 11

L20 NRL120 108 032 200 003 016 006 13 19

L22 Non-NRL 051 030 223 -029 013 011 -20 -17

L23 Non-NRL 047 009 200 -033 005 004 -23 -50

L24 NRL882 072 019 200 -014 010 006 -05 -07

L25 Non-NRL 120

008 000 000 18 85

L26 NRL882 023 009 200 -064 005 008 -54 -62

L27 Non-NRL 098 015 200 -001 008 003 09 23

L29 NRL120 090 008 173 -005 005 002 05 17

L30 NRL882 045 013 200 -035 007 006 -25 -38

L31 NRL882 080 038 223 -010 017 009 00 00

L32 Non-NRL 076 039 200 -012 020 011 -02 -02

L33 NRL-120 057 014 200 -024 007 005 -15 -26

L34 NRL120 061 016 200 -021 008 006 -12 -19

L35 NRL882 097 027 200 -001 014 006 08 13

L36 Non-NRL 090 030 200 -005 015 007 05 07

L37 NRL120 065 026 200 -019 013 009 -09 -10

L38 NRL120 093 014 237 -003 006 003 07 19

L39 Non-NRL 112

005 000 000 15 71

L40 NRL882 053 008 200 -028 004 003 -18 -46

L41 NRL120 080 003 200 -010 002 001 00 00

L42 NRL882 104 031 200 002 016 006 11 17

L44 NRL120 084 007 257 -008 003 001 02 09

L45 NRL882 065 018 200 -019 009 006 -09 -14

L46 NRL882 081 018 200 -009 009 005 01 01

L47 NRL882 141 027 173 015 016 005 25 47

L48 NRL882 059 028 200 -023 014 010 -13 -13

L49 Non-NRL 089 031 200 -005 016 008 05 06

L50 NRL120 073 005 200 -014 003 001 -04 -15

L51 Non-NRL 077 049 200 -011 025 014 -02 -01

L52 Non-NRL 091 012 200 -004 006 003 06 16

L53 Non-NRL 043 081 200 -037 041 041 -27 -07

L54 NRL882 074 035 200 -013 018 010 -03 -03

L55 Non-NRL 076 019 173 -012 011 006 -02 -03

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

3946

Annex 4 Participants quantitative performance

The z and ζ scores of all laboratories are reported in Tables A41 and A42 and in Figures

A41 and A42 for 40-3-2 soybean in T1 and T2 respectively For consistency the

reported results are shown with two decimals the performance scores with one decimal

The z scores are displayed in green (satisfactory) orange (questionable) or red cells

(unsatisfactory)

Table A41 Quantitative results (in mm ) and performance scores of participants of CT-0217 for 40-3-2 soybean in chicken feed (T1)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 064 008 200 -019 004 003 -10 -28

L02 NRL882 057 002 200 -024 001 001 -15 -67

L03 Non-NRL 073 1520 200 -014 760 452 -04 00

L04 Non-NRL 075

-012 000 000 -03 -13

L05 NRL120 073 007 257 -014 003 002 -04 -15

L06 NRL882 075 021 200 -012 011 006 -03 -04

L07 NRL882 065 013 200 -019 007 004 -09 -19

L08 NRL120 062 023 200 -021 012 008 -11 -13

L09 NRL882 081 027 200 -009 014 007 01 01

L12 Non-NRL 100 051 200 000 026 011 10 09

L13 NRL882 094 019 173 -003 011 005 07 13

L14 Non-NRL 063

-020 000 000 -10 -50

L15 NRL882 110 011 200 004 005 002 14 46

L16 Non-NRL 117

007 000 000 17 80

L17 NRL882 072 010 200 -014 005 003 -05 -12

L18 Non-NRL 040

-040 000 000 -30 -145

L19 Non-NRL 115 063 173 006 036 014 16 11

L20 NRL120 108 032 200 003 016 006 13 19

L22 Non-NRL 051 030 223 -029 013 011 -20 -17

L23 Non-NRL 047 009 200 -033 005 004 -23 -50

L24 NRL882 072 019 200 -014 010 006 -05 -07

L25 Non-NRL 120

008 000 000 18 85

L26 NRL882 023 009 200 -064 005 008 -54 -62

L27 Non-NRL 098 015 200 -001 008 003 09 23

L29 NRL120 090 008 173 -005 005 002 05 17

L30 NRL882 045 013 200 -035 007 006 -25 -38

L31 NRL882 080 038 223 -010 017 009 00 00

L32 Non-NRL 076 039 200 -012 020 011 -02 -02

L33 NRL-120 057 014 200 -024 007 005 -15 -26

L34 NRL120 061 016 200 -021 008 006 -12 -19

L35 NRL882 097 027 200 -001 014 006 08 13

L36 Non-NRL 090 030 200 -005 015 007 05 07

L37 NRL120 065 026 200 -019 013 009 -09 -10

L38 NRL120 093 014 237 -003 006 003 07 19

L39 Non-NRL 112

005 000 000 15 71

L40 NRL882 053 008 200 -028 004 003 -18 -46

L41 NRL120 080 003 200 -010 002 001 00 00

L42 NRL882 104 031 200 002 016 006 11 17

L44 NRL120 084 007 257 -008 003 001 02 09

L45 NRL882 065 018 200 -019 009 006 -09 -14

L46 NRL882 081 018 200 -009 009 005 01 01

L47 NRL882 141 027 173 015 016 005 25 47

L48 NRL882 059 028 200 -023 014 010 -13 -13

L49 Non-NRL 089 031 200 -005 016 008 05 06

L50 NRL120 073 005 200 -014 003 001 -04 -15

L51 Non-NRL 077 049 200 -011 025 014 -02 -01

L52 Non-NRL 091 012 200 -004 006 003 06 16

L53 Non-NRL 043 081 200 -037 041 041 -27 -07

L54 NRL882 074 035 200 -013 018 010 -03 -03

L55 Non-NRL 076 019 173 -012 011 006 -02 -03

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

4046

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L56 NRL882 115 022 173 006 013 005 16 30

L57 NRL120 075 020 200 -012 010 006 -03 -04

L58 NRL120 089 016 200 -005 008 004 05 11

L59 NRL882 137 038 200 014 019 006 23 37

L60 NRL882 068 017 200 -017 009 005 -07 -12

L61 Non-NRL 065

-019 000 000 -09 -43

L62 NRL120 130 037 237 011 016 005 21 38

L63 Non-NRL 088 006 200 -006 003 001 04 16

L64 NRL882 081 024 173 -009 014 007 01 01

L65 Non-NRL 116 065 200 006 033 012 16 13

L66 NRL882 048 024 200 -032 012 011 -22 -20

L68 NRL882 139 015 200 014 008 002 24 77

L69 Non-NRL 010 004 200 -101 002 008 -91 -105

L70 Non-NRL 057 010 200 -024 005 004 -15 -34

L72 NRL882 108 032 200 003 016 007 13 19

L74 NRL882 075 019 200 -012 010 005 -03 -05

L75 Non-NRL 862

094 000 000 103 499

L76 NRL882 099 038 200 000 019 008 09 11

L77 NRL120 075 009 226 -012 004 002 -03 -09

L78 NRL120 059 006 300 -023 002 001 -13 -52

L79 Non-NRL 046

-034 000 000 -24 -116

L80 NRL120 071 024 200 -015 012 007 -05 -07

L81 NRL882 071 028 200 -015 014 009 -05 -06

L82 NRL120 092 015 200 -004 008 004 06 15

L83 NRL120 114 013 173 006 008 003 15 44

L84 Non-NRL 071 030 200 -015 015 009 -05 -05

L85 NRL882 074 031 201 -013 015 009 -03 -04

L86 Non-NRL 066 022 200 -018 011 007 -08 -11

L87 NRL882 085 015 200 -007 008 004 03 06

L88 NRL-120 lt0044

L89 NRL882 079 022 200 -010 011 006 -01 -01

1 One NRL882 (L73) did not report a quantitative result for 40-3-2 in T1 because T1 is out of scope of the laboratory 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score 4 The less than value reported by the laboratory is considered unsatisfactory

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

4146

Table A42 Quantitative results (in mm ) and performance scores of participants of CT-0217 for

40-3-2 soybean in soybean flour (T2)

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L01 NRL120 071 007 200 -015 004 002 -03 -12

L02 NRL882 081 003 200 -009 002 001 03 19

L03 Non-NRL 088 1520 200 -006 760 375 06 00

L04 Non-NRL 089

-005 000 000 07 56

L05 NRL120 075 005 257 -012 002 001 -01 -04

L06 NRL882 080 022 200 -010 011 006 02 04

L07 NRL882 076 009 200 -012 005 003 00 00

L08 NRL120 042 016 200 -038 008 008 -26 -31

L09 NRL882 081 027 200 -009 014 007 03 04

L12 Non-NRL 076 039 200 -012 020 011 00 00

L13 NRL882 069 011 173 -016 006 004 -04 -10

L14 Non-NRL 075

-012 000 000 -01 -05

L15 NRL882 084 011 200 -008 005 003 04 14

L16 Non-NRL 088

-006 000 000 06 52

L17 NRL882 075 010 200 -012 005 003 -01 -02

L18 Non-NRL 034

-047 000 000 -35 -287

L19 Non-NRL 089 016 173 -005 009 005 07 15

L20 NRL120 104 031 200 002 016 006 14 21

L22 Non-NRL 081 026 223 -009 012 006 03 04

L23 Non-NRL 064 012 200 -019 006 004 -08 -18

L24 NRL882 079 021 200 -010 011 006 02 03

L25 Non-NRL 085

-007 000 000 05 39

L26 NRL882 117 039 200 007 020 007 19 25

L27 Non-NRL 118 018 200 007 009 003 19 54

L29 NRL120 078 004 173 -011 002 001 01 06

L30 NRL882 069 020 200 -016 010 006 -04 -07

L31 NRL882 096 077 233 -002 033 015 10 07

L32 Non-NRL 090 039 200 -005 020 009 07 08

L33 NRL-120 078 004 200 -011 002 001 01 06

L34 NRL120 073 014 200 -014 007 004 -02 -04

L35 NRL882 095 027 200 -002 014 006 10 15

L36 Non-NRL 077 003 200 -011 002 001 01 03

L37 NRL120 075 010 200 -012 005 003 -01 -02

L38 NRL120 119 008 237 008 003 001 19 112

L39 Non-NRL 062

-021 000 000 -09 -73

L40 NRL882 045 008 200 -035 004 004 -23 -56

L41 NRL120 087 002 200 -006 001 000 06 44

L42 NRL882 068 020 200 -017 010 006 -05 -08

L43 Non-NRL 065

-019 000 000 -07 -56

L44 NRL120 074 006 220 -013 003 002 -01 -06

L45 NRL882 073 021 200 -014 011 006 -02 -03

L46 NRL882 081 018 200 -009 009 005 03 05

L47 NRL882 062 012 173 -021 007 005 -09 -18

L48 NRL882 074 035 173 -013 020 012 -01 -01

L49 Non-NRL 073 026 200 -014 013 008 -02 -02

L50 NRL120 081 007 200 -009 004 002 03 12

L51 Non-NRL 068 044 200 -017 022 014 -05 -03

L52 Non-NRL 073 011 200 -014 006 003 -02 -05

L53 Non-NRL 058 081 200 -024 041 030 -12 -04

L54 NRL882 073 026 200 -014 013 008 -02 -02

L55 Non-NRL 072 018 173 -014 010 006 -02 -04

L56 NRL882 066 018 173 -018 010 007 -06 -09

L57 NRL120 060 020 200 -022 010 007 -10 -14

L58 NRL120 067 011 200 -017 006 004 -06 -15

L59 NRL882 076 022 200 -012 011 006 00 00

L60 NRL882 080 020 200 -010 010 005 02 04

L61 Non-NRL 050

-030 000 000 -18 -150

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report

4246

Lab1 Category xi U k2 Log(xi) u(xi) u(xi-log) z score ζ score3

L62 NRL120 075 012 245 -012 005 003 -01 -02

L63 Non-NRL 075 004 200 -012 002 001 -01 -04

L64 NRL882 078 023 173 -011 013 007 01 01

L65 Non-NRL 051 023 200 -029 012 010 -17 -18

L66 NRL882 073 024 200 -014 012 007 -02 -02

L68 NRL882 091 011 200 -004 006 003 08 27

L69 Non-NRL 092 035 200 -004 018 008 08 10

L70 Non-NRL 050 009 200 -030 005 004 -18 -45

L72 NRL882 092 028 200 -004 014 007 08 12

L73 NRL882 092 030 200 -004 015 007 08 11

L74 NRL882 064 013 200 -019 007 004 -08 -16

L75 Non-NRL 644

081 000 000 93 760

L76 NRL882 056 099 200 -025 050 038 -13 -03

L77 NRL120 054 006 226 -027 003 002 -15 -61

L78 NRL120 065 001 300 -019 000 000 -07 -55

L79 Non-NRL 071

-015 000 000 -03 -25

L80 NRL120 095 030 200 -002 015 007 10 14

L81 NRL882 077 032 200 -011 016 009 01 01

L82 NRL120 102 032 200 001 016 007 13 18

L83 NRL120 084 005 173 -008 003 001 04 22

L84 Non-NRL 045 030 200 -035 015 014 -23 -16

L85 NRL882 082 015 212 -009 007 004 03 08

L86 Non-NRL 073 024 200 -014 012 007 -02 -02

L87 NRL882 077 012 200 -011 006 003 01 01

L88 NRL120 065 020 200 -019 010 007 -07 -10

L89 NRL882 077 021 200 -011 011 006 01 01

1 All participating NRL882 reported a quantitative result for 40-3-2 soybean in T2 2 If the k factor was not reported by the laboratory a value of 173 was assigned (radic3) for calculation of the ζ score 3 The ζ scores are provided for information only with indicative of a questionable score and of an unsatisfactory score

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report 4346

Figure A41 Laboratory results for soybean event 40-3-2 in test item 1 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

EURL GMFF CT 0217 report 4446

Figure A42 Laboratory results for soybean event 40-3-2 in test item 2 and kernel density distribution (insert)

The horizontal full line shows the assigned value (on the log scale) the dashed black lines represent the expanded measurement uncertainty of the assigned value and the wider interval (dashed red lines) represents the limits of satisfaction (|z|le 20) Laboratory results are shown with the expanded measurement uncertainties (when reported) the expanded measurement uncertainty of L03 is out of scale

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number ()

00 800 6 7 8 9 10 11 () The information given is free as are most calls (though some operators phone boxes or hotels may

charge you)

More information on the European Union is available on the internet (httpeuropaeu)

HOW TO OBTAIN EU PUBLICATIONS

Free publications

bull one copy

via EU Bookshop (httpbookshopeuropaeu)

bull more than one copy or postersmaps

from the European Unionrsquos representations (httpeceuropaeurepresent_enhtm) from the delegations in non-EU countries (httpeeaseuropaeudelegationsindex_enhtm) by contacting the Europe Direct service (httpeuropaeueuropedirectindex_enhtm) or calling 00 800 6 7 8 9 10 11 (freephone number from anywhere in the EU) () () The information given is free as are most calls (though some operators phone boxes or hotels may charge you)

Priced publications

bull via EU Bookshop (httpbookshopeuropaeu)

KJ-N

A-2

8873-E

N-N

doi 102760461562

ISBN 978-92-79-76303-8